Articoli di riviste sul tema "Benign overfitting"
Segui questo link per vedere altri tipi di pubblicazioni sul tema: Benign overfitting.
Cita una fonte nei formati APA, MLA, Chicago, Harvard e in molti altri stili
Vedi i top-46 articoli di riviste per l'attività di ricerca sul tema "Benign overfitting".
Accanto a ogni fonte nell'elenco di riferimenti c'è un pulsante "Aggiungi alla bibliografia". Premilo e genereremo automaticamente la citazione bibliografica dell'opera scelta nello stile citazionale di cui hai bisogno: APA, MLA, Harvard, Chicago, Vancouver ecc.
Puoi anche scaricare il testo completo della pubblicazione scientifica nel formato .pdf e leggere online l'abstract (il sommario) dell'opera se è presente nei metadati.
Vedi gli articoli di riviste di molte aree scientifiche e compila una bibliografia corretta.
1
Bartlett, Peter L., Philip M. Long, Gábor Lugosi e Alexander Tsigler. "Benign overfitting in linear regression". Proceedings of the National Academy of Sciences 117, n. 48 (24 aprile 2020): 30063–70. http://dx.doi.org/10.1073/pnas.1907378117.
Abstract (sommario):
The phenomenon of benign overfitting is one of the key mysteries uncovered by deep learning methodology: deep neural networks seem to predict well, even with a perfect fit to noisy training data. Motivated by this phenomenon, we consider when a perfect fit to training data in linear regression is compatible with accurate prediction. We give a characterization of linear regression problems for which the minimum norm interpolating prediction rule has near-optimal prediction accuracy. The characterization is in terms of two notions of the effective rank of the data covariance. It shows that overparameterization is essential for benign overfitting in this setting: the number of directions in parameter space that are unimportant for prediction must significantly exceed the sample size. By studying examples of data covariance properties that this characterization shows are required for benign overfitting, we find an important role for finite-dimensional data: the accuracy of the minimum norm interpolating prediction rule approaches the best possible accuracy for a much narrower range of properties of the data distribution when the data lie in an infinite-dimensional space vs. when the data lie in a finite-dimensional space with dimension that grows faster than the sample size.
2
Peters, Evan, e Maria Schuld. "Generalization despite overfitting in quantum machine learning models". Quantum 7 (20 dicembre 2023): 1210. http://dx.doi.org/10.22331/q-2023-12-20-1210.
Abstract (sommario):
The widespread success of deep neural networks has revealed a surprise in classical machine learning: very complex models often generalize well while simultaneously overfitting training data. This phenomenon of benign overfitting has been studied for a variety of classical models with the goal of better understanding the mechanisms behind deep learning. Characterizing the phenomenon in the context of quantum machine learning might similarly improve our understanding of the relationship between overfitting, overparameterization, and generalization. In this work, we provide a characterization of benign overfitting in quantum models. To do this, we derive the behavior of a classical interpolating Fourier features models for regression on noisy signals, and show how a class of quantum models exhibits analogous features, thereby linking the structure of quantum circuits (such as data-encoding and state preparation operations) to overparameterization and overfitting in quantum models. We intuitively explain these features according to the ability of the quantum model to interpolate noisy data with locally "spiky" behavior and provide a concrete demonstration example of benign overfitting.
3
Bartlett, Peter L., Andrea Montanari e Alexander Rakhlin. "Deep learning: a statistical viewpoint". Acta Numerica 30 (maggio 2021): 87–201. http://dx.doi.org/10.1017/s0962492921000027.
Abstract (sommario):
The remarkable practical success of deep learning has revealed some major surprises from a theoretical perspective. In particular, simple gradient methods easily find near-optimal solutions to non-convex optimization problems, and despite giving a near-perfect fit to training data without any explicit effort to control model complexity, these methods exhibit excellent predictive accuracy. We conjecture that specific principles underlie these phenomena: that overparametrization allows gradient methods to find interpolating solutions, that these methods implicitly impose regularization, and that overparametrization leads to benign overfitting, that is, accurate predictions despite overfitting training data. In this article, we survey recent progress in statistical learning theory that provides examples illustrating these principles in simpler settings. We first review classical uniform convergence results and why they fall short of explaining aspects of the behaviour of deep learning methods. We give examples of implicit regularization in simple settings, where gradient methods lead to minimal norm functions that perfectly fit the training data. Then we review prediction methods that exhibit benign overfitting, focusing on regression problems with quadratic loss. For these methods, we can decompose the prediction rule into a simple component that is useful for prediction and a spiky component that is useful for overfitting but, in a favourable setting, does not harm prediction accuracy. We focus specifically on the linear regime for neural networks, where the network can be approximated by a linear model. In this regime, we demonstrate the success of gradient flow, and we consider benign overfitting with two-layer networks, giving an exact asymptotic analysis that precisely demonstrates the impact of overparametrization. We conclude by highlighting the key challenges that arise in extending these insights to realistic deep learning settings.
4
Wang, Ke, e Christos Thrampoulidis. "Binary Classification of Gaussian Mixtures: Abundance of Support Vectors, Benign Overfitting, and Regularization". SIAM Journal on Mathematics of Data Science 4, n. 1 (marzo 2022): 260–84. http://dx.doi.org/10.1137/21m1415121.
5
Hu, Wei. "Understanding Surprising Generalization Phenomena in Deep Learning". Proceedings of the AAAI Conference on Artificial Intelligence 38, n. 20 (24 marzo 2024): 22669. http://dx.doi.org/10.1609/aaai.v38i20.30285.
Abstract (sommario):
Deep learning has exhibited a number of surprising generalization phenomena that are not captured by classical statistical learning theory. This talk will survey some of my work on the theoretical characterizations of several such intriguing phenomena: (1) Implicit regularization: A major mystery in deep learning is that deep neural networks can often generalize well despite their excessive expressive capacity. Towards explaining this mystery, it has been suggested that commonly used gradient-based optimization algorithms enforce certain implicit regularization which effectively constrains the model capacity. (2) Benign overfitting: In certain scenarios, a model can perfectly fit noisily labeled training data, but still archives near-optimal test error at the same time, which is very different from the classical notion of overfitting. (3) Grokking: In certain scenarios, a model initially achieves perfect training accuracy but no generalization (i.e. no better than a random predictor), and upon further training, transitions to almost perfect generalization. Theoretically establishing these properties often involves making appropriate high-dimensional assumptions on the problem as well as a careful analysis of the training dynamics.
6
Montaha, Sidratul, Sami Azam, A. K. M. Rakibul Haque Rafid, Sayma Islam, Pronab Ghosh e Mirjam Jonkman. "A shallow deep learning approach to classify skin cancer using down-scaling method to minimize time and space complexity". PLOS ONE 17, n. 8 (4 agosto 2022): e0269826. http://dx.doi.org/10.1371/journal.pone.0269826.
Abstract (sommario):
The complex feature characteristics and low contrast of cancer lesions, a high degree of inter-class resemblance between malignant and benign lesions, and the presence of various artifacts including hairs make automated melanoma recognition in dermoscopy images quite challenging. To date, various computer-aided solutions have been proposed to identify and classify skin cancer. In this paper, a deep learning model with a shallow architecture is proposed to classify the lesions into benign and malignant. To achieve effective training while limiting overfitting problems due to limited training data, image preprocessing and data augmentation processes are introduced. After this, the ‘box blur’ down-scaling method is employed, which adds efficiency to our study by reducing the overall training time and space complexity significantly. Our proposed shallow convolutional neural network (SCNN_12) model is trained and evaluated on the Kaggle skin cancer data ISIC archive which was augmented to 16485 images by implementing different augmentation techniques. The model was able to achieve an accuracy of 98.87% with optimizer Adam and a learning rate of 0.001. In this regard, parameter and hyper-parameters of the model are determined by performing ablation studies. To assert no occurrence of overfitting, experiments are carried out exploring k-fold cross-validation and different dataset split ratios. Furthermore, to affirm the robustness the model is evaluated on noisy data to examine the performance when the image quality gets corrupted.This research corroborates that effective training for medical image analysis, addressing training time and space complexity, is possible even with a lightweighted network using a limited amount of training data.
7
Windisch, Paul, Carole Koechli, Susanne Rogers, Christina Schröder, Robert Förster, Daniel R. Zwahlen e Stephan Bodis. "Machine Learning for the Detection and Segmentation of Benign Tumors of the Central Nervous System: A Systematic Review". Cancers 14, n. 11 (27 maggio 2022): 2676. http://dx.doi.org/10.3390/cancers14112676.
Abstract (sommario):
Objectives: To summarize the available literature on using machine learning (ML) for the detection and segmentation of benign tumors of the central nervous system (CNS) and to assess the adherence of published ML/diagnostic accuracy studies to best practice. Methods: The MEDLINE database was searched for the use of ML in patients with any benign tumor of the CNS, and the records were screened according to PRISMA guidelines. Results: Eleven retrospective studies focusing on meningioma (n = 4), vestibular schwannoma (n = 4), pituitary adenoma (n = 2) and spinal schwannoma (n = 1) were included. The majority of studies attempted segmentation. Links to repositories containing code were provided in two manuscripts, and no manuscripts shared imaging data. Only one study used an external test set, which raises the question as to whether some of the good performances that have been reported were caused by overfitting and may not generalize to data from other institutions. Conclusion: Using ML for detecting and segmenting benign brain tumors is still in its infancy. Stronger adherence to ML best practices could facilitate easier comparisons between studies and contribute to the development of models that are more likely to one day be used in clinical practice.
8
Liang, ShuFen, HuiLin Liu, FangChen Yang, Chuanbo Qin e Yue Feng. "Classification of Benign and Malignant Pulmonary Nodules Using a Regularized Extreme Learning Machine". Journal of Medical Imaging and Health Informatics 11, n. 8 (1 agosto 2021): 2117–23. http://dx.doi.org/10.1166/jmihi.2021.3448.
Abstract (sommario):
An L1/L2-norm-bound extreme learning machine classification algorithm is proposed to improve the accuracy of distinguishing between benign and malignant pulmonary nodules. In this algorithm, features extracted from the segmented lung nodule using the histogram of oriented gradients method are used as inputs. L1-norm can promote sparsity in the weights of the output layer, and L2-norm can smooth output weights. The combination of the L1 norm and L2 norm can simplify the complexity of the network and prevent overfitting to improve classification accuracy. For each newly tested lung nodule, the algorithm outputs a class label of either benign or malignant. The accuracy, sensitivity, and specificity reached 94.12%, 93%, and 95% respectively over the lung image database consortium and image database resource initiative dataset. Compared with other algorithms, the average values of the three metrics increased by 6.5%, 7.94%, and 4.32%, respectively. An accuracy score of 95.83% can be achieved over a set of 120 urinary sediment images. Therefore, this algorithm has a good classification effect of pulmonary nodules.
9
Liu, Xinwei, Xiaojun Jia, Jindong Gu, Yuan Xun, Siyuan Liang e Xiaochun Cao. "Does Few-Shot Learning Suffer from Backdoor Attacks?" Proceedings of the AAAI Conference on Artificial Intelligence 38, n. 18 (24 marzo 2024): 19893–901. http://dx.doi.org/10.1609/aaai.v38i18.29965.
Abstract (sommario):
The field of few-shot learning (FSL) has shown promising results in scenarios where training data is limited, but its vulnerability to backdoor attacks remains largely unexplored. We first explore this topic by first evaluating the performance of the existing backdoor attack methods on few-shot learning scenarios. Unlike in standard supervised learning, existing backdoor attack methods failed to perform an effective attack in FSL due to two main issues. Firstly, the model tends to overfit to either benign features or trigger features, causing a tough trade-off between attack success rate and benign accuracy. Secondly, due to the small number of training samples, the dirty label or visible trigger in the support set can be easily detected by victims, which reduces the stealthiness of attacks. It seemed that FSL could survive from backdoor attacks. However, in this paper, we propose the Few-shot Learning Backdoor Attack (FLBA) to show that FSL can still be vulnerable to backdoor attacks. Specifically, we first generate a trigger to maximize the gap between poisoned and benign features. It enables the model to learn both benign and trigger features, which solves the problem of overfitting. To make it more stealthy, we hide the trigger by optimizing two types of imperceptible perturbation, namely attractive and repulsive perturbation, instead of attaching the trigger directly. Once we obtain the perturbations, we can poison all samples in the benign support set into a hidden poisoned support set and fine-tune the model on it. Our method demonstrates a high Attack Success Rate (ASR) in FSL tasks with different few-shot learning paradigms while preserving clean accuracy and maintaining stealthiness. This study reveals that few-shot learning still suffers from backdoor attacks, and its security should be given attention.
10
Doimo, Diego, Aldo Glielmo, Sebastian Goldt e Alessandro Laio. "Redundant representations help generalization in wide neural networks * , †". Journal of Statistical Mechanics: Theory and Experiment 2023, n. 11 (1 novembre 2023): 114011. http://dx.doi.org/10.1088/1742-5468/aceb4f.
Abstract (sommario):
Abstract Deep neural networks (DNNs) defy the classical bias-variance trade-off; adding parameters to a DNN that interpolates its training data will typically improve its generalization performance. Explaining the mechanism behind this ‘benign overfitting’ in deep networks remains an outstanding challenge. Here, we study the last hidden layer representations of various state-of-the-art convolutional neural networks and find that if the last hidden representation is wide enough, its neurons tend to split into groups that carry identical information and differ from each other only by statistically independent noise. The number of these groups increases linearly with the width of the layer, but only if the width is above a critical value. We show that redundant neurons appear only when the training is regularized and the training error is zero.
11
Li, Jian, Yong Liu e Weiping Wang. "High-Dimensional Analysis for Generalized Nonlinear Regression: From Asymptotics to Algorithm". Proceedings of the AAAI Conference on Artificial Intelligence 38, n. 12 (24 marzo 2024): 13500–13508. http://dx.doi.org/10.1609/aaai.v38i12.29253.
Abstract (sommario):
Overparameterization often leads to benign overfitting, where deep neural networks can be trained to overfit the training data but still generalize well on unseen data. However, it lacks a generalized asymptotic framework for nonlinear regressions and connections to conventional complexity notions. In this paper, we propose a generalized high-dimensional analysis for nonlinear regression models, including various nonlinear feature mapping methods and subsampling. Specifically, we first provide an implicit regularization parameter and asymptotic equivalents related to a classical complexity notion, i.e., effective dimension. We then present a high-dimensional analysis for nonlinear ridge regression and extend it to ridgeless regression in the under-parameterized and over-parameterized regimes, respectively. We find that the limiting risks decrease with the effective dimension. Motivated by these theoretical findings, we propose an algorithm, namely RFRed, to improve generalization ability. Finally, we validate our theoretical findings and the proposed algorithm through several experiments.
12
Alkhaleefah, Mohammad, Shang-Chih Ma, Yang-Lang Chang, Bormin Huang, Praveen Kumar Chittem e Vishnu Priya Achhannagari. "Double-Shot Transfer Learning for Breast Cancer Classification from X-Ray Images". Applied Sciences 10, n. 11 (9 giugno 2020): 3999. http://dx.doi.org/10.3390/app10113999.
Abstract (sommario):
Differentiation between benign and malignant breast cancer cases in X-ray images can be difficult due to their similar features. In recent studies, the transfer learning technique has been used to classify benign and malignant breast cancer by fine-tuning various pre-trained networks such as AlexNet, visual geometry group (VGG), GoogLeNet, and residual network (ResNet) on breast cancer datasets. However, these pre-trained networks have been trained on large benchmark datasets such as ImageNet, which do not contain labeled images related to breast cancers which lead to poor performance. In this research, we introduce a novel technique based on the concept of transfer learning, called double-shot transfer learning (DSTL). DSTL is used to improve the overall accuracy and performance of the pre-trained networks for breast cancer classification. DSTL updates the learnable parameters (weights and biases) of any pre-trained network by fine-tuning them on a large dataset that is similar to the target dataset. Then, the updated networks are fine-tuned with the target dataset. Moreover, the number of X-ray images is enlarged by a combination of augmentation methods including different variations of rotation, brightness, flipping, and contrast to reduce overfitting and produce robust results. The proposed approach has demonstrated a significant improvement in classification accuracy and performance of the pre-trained networks, making them more suitable for medical imaging.
13
Istighosah, Maie, Andi Sunyoto e Tonny Hidayat. "Breast Cancer Detection in Histopathology Images using ResNet101 Architecture". sinkron 8, n. 4 (1 ottobre 2023): 2138–49. http://dx.doi.org/10.33395/sinkron.v8i4.12948.
Abstract (sommario):
Cancer is a significant challenge in many fields, especially health and medicine. Breast cancer is among the most common and frequent cancers in women worldwide. Early detection of cancer is the main step for early treatment and increasing the chances of patient survival. As the convolutional neural network method has grown in popularity, breast cancer can be easily identified without the help of experts. Using BreaKHis histopathology data, this project will assess the efficacy of the CNN architecture ResNet101 for breast cancer image classification. The dataset is divided into two classes, namely 1146 malignant and 547 benign. The treatment of data preprocessing is considered. The implementation of data augmentation in the benign class to obtain data balance between the two classes and prevent overfitting. The BreaKHis dataset has noise and uneven color distribution. Approaches such as bilateral filtering, image enhancement, and color normalization were chosen to enhance image quality. Adding flatten, dense, and dropout layers to the ResNet101 architecture is applied to improve the model performance. Parameters were modified during the training stage to achieve optimal model performance. The Adam optimizer was used with a learning rate 0.0001 and a batch size of 32. Furthermore, the model was trained for 100 epochs. The accuracy, precision, recall, and f1-score results are 98.7%, 98.73%, 98.7%, and 98.7%, respectively. According to the results, the proposed ResNet101 model outperforms the standard technique as well as other architectures.
14
Anjum, Sunila, Imran Ahmed, Muhammad Asif, Hanan Aljuaid, Fahad Alturise, Yazeed Yasin Ghadi e Rashad Elhabob. "Lung Cancer Classification in Histopathology Images Using Multiresolution Efficient Nets". Computational Intelligence and Neuroscience 2023 (16 ottobre 2023): 1–12. http://dx.doi.org/10.1155/2023/7282944.
Abstract (sommario):
Histopathological images are very effective for investigating the status of various biological structures and diagnosing diseases like cancer. In addition, digital histopathology increases diagnosis precision and provides better image quality and more detail for the pathologist with multiple viewing options and team annotations. As a result of the benefits above, faster treatment is available, increasing therapy success rates and patient recovery and survival chances. However, the present manual examination of these images is tedious and time-consuming for pathologists. Therefore, reliable automated techniques are needed to effectively classify normal and malignant cancer images. This paper applied a deep learning approach, namely, EfficientNet and its variants from B0 to B7. We used different image resolutions for each model, from 224 × 224 pixels to 600 × 600 pixels. We also applied transfer learning and parameter tuning techniques to improve the results and overcome the overfitting problem. We collected the dataset from the Lung and Colon Cancer Histopathological Image LC25000 image dataset. The dataset acquisition consists of 25,000 histopathology images of five classes (lung adenocarcinoma, lung squamous cell carcinoma, benign lung tissue, colon adenocarcinoma, and colon benign tissue). Then, we performed preprocessing on the dataset to remove the noisy images and bring them into a standard format. The model’s performance was evaluated in terms of classification accuracy and loss. We have achieved good accuracy results for all variants; however, the results of EfficientNetB2 stand excellent, with an accuracy of 97% for 260 × 260 pixels resolution images.
15
Nadkarni, Swati, e Kevin Noronha. "Breast cancer detection using ensemble of convolutional neural networks". International Journal of Electrical and Computer Engineering (IJECE) 14, n. 1 (1 febbraio 2024): 1041. http://dx.doi.org/10.11591/ijece.v14i1.pp1041-1047.
Abstract (sommario):
Early detection leading to timely treatment in the initial stages of cancer may decrease the breast cancer death rate. We propose deep learning techniques along with image processing for the detection of tumors. The availability of online datasets and advances in graphical processing units (GPU) have promoted the application of deep learning models for the detection of breast cancer. In this paper, deep learning models using convolutional neural network (CNN) have been built to automatically classify mammograms into benign and malignant. Issues like overfitting and dataset imbalance are overcome. Experimentation has been done on two publicly available datasets, namely mammographic image analysis society (MIAS) database and digital database for screening mammography (DDSM). Robustness of the models is accomplished by merging the datasets. In our experimentation, MatConvNet has achieved an accuracy of 94.2% on the merged dataset, performing the best amongst all the CNN models used individually. Hungarian optimization algorithm is employed for selection of individual CNN models to form an ensemble. Ensemble of CNN models led to an improved performance, resulting in an accuracy of 95.7%.
16
Ren, Cheng, e Shouming Hou. "A Hybrid Deep Learning Approach for Lung Nodule Classification". Frontiers in Computing and Intelligent Systems 8, n. 1 (10 maggio 2024): 6–12. http://dx.doi.org/10.54097/498fxm65.
Abstract (sommario):
Lung cancer has the highest morbidity and mortality rates worldwide. Pulmonary nodules are an early manifestation of lung cancer. Therefore, accurate classification of pulmonary nodules is of great significance for the early diagnosis and treatment of lung cancer. However, the classification of lung nodules is a complex and time-consuming task requiring extensive image reading and analysis by expert radiologists. Therefore, using deep learning technology to assist doctors in detecting and classifying pulmonary nodules has become a current research trend. A lightweight classification model named Res-VGG is proposed for classifying lung nodules as benign or malignant. The Res-VGG model improves on VGG16 by reducing the use of convolutional and fully connected layers. To reduce overfitting, residual connections are introduced. The training of the model was performed on the LUNA16 database, and a ten-fold cross-validation method was used to evaluate the performance of the model. In addition, the Res-VGG model was compared with three other common classification networks, and the results showed that the Res-VGG model outperformed the other models in terms of accuracy, sensitivity, and specificity.
17
Zi Wei, Yee, Marina Md-Arshad, Adlina Abdul Samad e Norafida Ithnin. "Comparing Malware Attack Detection using Machine Learning Techniques in IoT Network Traffic". International Journal of Innovative Computing 13, n. 1 (30 maggio 2023): 21–27. http://dx.doi.org/10.11113/ijic.v13n1.384.
Abstract (sommario):
Most IoT devices are designed and built for cheap and basic functions, therefore, the security aspects of these devices are not taken seriously. Yet, IoT devices tend to play an important role in this era, where the amount of IoT devices is predicted to exceed the number of traditional computing devices such as desktops and laptops. This causes more and more cybersecurity attacks to target IoT devices and malware attack is known to be the most common attack in IoT networks. However, most research only focuses on malware detection in traditional computing devices. The purpose of this research is to compare the performance of Random Forest and Naïve Bayes algorithm in terms of accuracy, precision, recall and F1-score in classifying the malware attack and benign traffic in IoT network traffic. Research is conducted with the Aposemat IoT-23 dataset, a labelled dataset that contains IoT malware infection traffic and IoT benign traffic. To determine the data in IoT network traffic packets that are useful for threat detection, a study is conducted and the threat data is cleaned up and prepared using RStudio and RapidMiner Studio. Random Forest and Naïve Bayes algorithm is used to train and classify the cleaned dataset. Random Forest can prevent the model from overfitting while Naïve Bayes requires less training time. Lastly, the accuracy, precision, recall and F1-score of the machine learning algorithms are compared and discussed. The research result displays the Random Forest as the best machine learning algorithm in classifying the malware attack traffic.
18
Palla, Tarun Ganesh, e Shahab Tayeb. "Intelligent Mirai Malware Detection for IoT Nodes". Electronics 10, n. 11 (24 maggio 2021): 1241. http://dx.doi.org/10.3390/electronics10111241.
Abstract (sommario):
The advancement in recent IoT devices has led to catastrophic attacks on the devices resulting in breaches in user privacy and exhausting resources of various organizations, so that users and organizations expend increased time and money. One such harmful malware is Mirai, which has created worldwide recognition by impacting the digital world. There are several ways to detect Mirai, but the Machine Learning approach has proved to be accurate and reliable in detecting malware. In this research, a novel-based approach of detecting Mirai using Machine Learning Algorithm is proposed and implemented in Matlab and Python. To evaluate the proposed approaches, Mirai and Benign datasets are considered and training is performed on the dataset comprised of a Training set, Cross-Validation set and Test set using Artificial Neural Network (ANN) consisting of neurons in the hidden layer, which provides consistent accuracy, precision, recall and F-1 score. In this research, an accurate number of hidden layers and neurons are chosen to avoid the problem of Overfitting. This research provides a comparative analysis between ANN and Random Forest models of the dataset formed by merging Mirai and benign datasets of the Mirai malware detection pertaining to seven IoT devices. The dataset used in this research is “N-BaIoT” dataset, which represents data in the features infected by Mirai Malware. The results are found to be accurate and reliable as the best performance was achieved with an accuracy of 92.8% and False Negative rate of 0.3% and F-1 score of 0.99. The expected outcomes of this project, include major findings towards cost-effective Learning solutions in detecting Mirai Malware strains.
19
Alruwaili, Madallah, e Walaa Gouda. "Automated Breast Cancer Detection Models Based on Transfer Learning". Sensors 22, n. 3 (24 gennaio 2022): 876. http://dx.doi.org/10.3390/s22030876.
Abstract (sommario):
Breast cancer is among the leading causes of mortality for females across the planet. It is essential for the well-being of women to develop early detection and diagnosis techniques. In mammography, focus has contributed to the use of deep learning (DL) models, which have been utilized by radiologists to enhance the needed processes to overcome the shortcomings of human observers. The transfer learning method is being used to distinguish malignant and benign breast cancer by fine-tuning multiple pre-trained models. In this study, we introduce a framework focused on the principle of transfer learning. In addition, a mixture of augmentation strategies were used to prevent overfitting and produce stable outcomes by increasing the number of mammographic images; including several rotation combinations, scaling, and shifting. On the Mammographic Image Analysis Society (MIAS) dataset, the proposed system was evaluated and achieved an accuracy of 89.5% using (residual network-50) ResNet50, and achieved an accuracy of 70% using the Nasnet-Mobile network. The proposed system demonstrated that pre-trained classification networks are significantly more effective and efficient, making them more acceptable for medical imaging, particularly for small training datasets.
20
Liu, Yaning, Lin Han, Hexiang Wang e Bo Yin. "Classification of papillary thyroid carcinoma histological images based on deep learning". Journal of Intelligent & Fuzzy Systems 40, n. 6 (21 giugno 2021): 12011–21. http://dx.doi.org/10.3233/jifs-210100.
Abstract (sommario):
Papillary thyroid carcinoma (PTC) is a common carcinoma in thyroid. As many benign thyroid nodules have the papillary structure which could easily be confused with PTC in morphology. Thus, pathologists have to take a lot of time on differential diagnosis of PTC besides personal diagnostic experience and there is no doubt that it is subjective and difficult to obtain consistency among observers. To address this issue, we applied deep learning to the differential diagnosis of PTC and proposed a histological image classification method for PTC based on the Inception Residual convolutional neural network (IRCNN) and support vector machine (SVM). First, in order to expand the dataset and solve the problem of histological image color inconsistency, a pre-processing module was constructed that included color transfer and mirror transform. Then, to alleviate overfitting of the deep learning model, we optimized the convolution neural network by combining Inception Network and Residual Network to extract image features. Finally, the SVM was trained via image features extracted by IRCNN to perform the classification task. Experimental results show effectiveness of the proposed method in the classification of PTC histological images.
21
Ullah, Naeem, Ali Javed, Ali Alhazmi, Syed M. Hasnain, Ali Tahir e Rehan Ashraf. "TumorDetNet: A unified deep learning model for brain tumor detection and classification". PLOS ONE 18, n. 9 (27 settembre 2023): e0291200. http://dx.doi.org/10.1371/journal.pone.0291200.
Abstract (sommario):
Accurate diagnosis of the brain tumor type at an earlier stage is crucial for the treatment process and helps to save the lives of a large number of people worldwide. Because they are non-invasive and spare patients from having an unpleasant biopsy, magnetic resonance imaging (MRI) scans are frequently employed to identify tumors. The manual identification of tumors is difficult and requires considerable time due to the large number of three-dimensional images that an MRI scan of one patient’s brain produces from various angles. Moreover, the variations in location, size, and shape of the brain tumor also make it challenging to detect and classify different types of tumors. Thus, computer-aided diagnostics (CAD) systems have been proposed for the detection of brain tumors. In this paper, we proposed a novel unified end-to-end deep learning model named TumorDetNet for brain tumor detection and classification. Our TumorDetNet framework employs 48 convolution layers with leaky ReLU (LReLU) and ReLU activation functions to compute the most distinctive deep feature maps. Moreover, average pooling and a dropout layer are also used to learn distinctive patterns and reduce overfitting. Finally, one fully connected and a softmax layer are employed to detect and classify the brain tumor into multiple types. We assessed the performance of our method on six standard Kaggle brain tumor MRI datasets for brain tumor detection and classification into (malignant and benign), and (glioma, pituitary, and meningioma). Our model successfully identified brain tumors with remarkable accuracy of 99.83%, classified benign and malignant brain tumors with an ideal accuracy of 100%, and meningiomas, pituitary, and gliomas tumors with an accuracy of 99.27%. These outcomes demonstrate the potency of the suggested methodology for the reliable identification and categorization of brain tumors.
22
Zawad, Syed, Ahsan Ali, Pin-Yu Chen, Ali Anwar, Yi Zhou, Nathalie Baracaldo, Yuan Tian e Feng Yan. "Curse or Redemption? How Data Heterogeneity Affects the Robustness of Federated Learning". Proceedings of the AAAI Conference on Artificial Intelligence 35, n. 12 (18 maggio 2021): 10807–14. http://dx.doi.org/10.1609/aaai.v35i12.17291.
Abstract (sommario):
Data heterogeneity has been identified as one of the key features in federated learning but often overlooked in the lens of robustness to adversarial attacks. This paper focuses on characterizing and understanding its impact on backdooring attacks in federated learning through comprehensive experiments using synthetic and the LEAF benchmarks. The initial impression driven by our experimental results suggests that data heterogeneity is the dominant factor in the effectiveness of attacks and it may be a redemption for defending against backdooring as it makes the attack less efficient, more challenging to design effective attack strategies, and the attack result also becomes less predictable. However, with further investigations, we found data heterogeneity is more of a curse than a redemption as the attack effectiveness can be significantly boosted by simply adjusting the client-side backdooring timing. More importantly, data heterogeneity may result in overfitting at the local training of benign clients, which can be utilized by attackers to disguise themselves and fool skewed-feature based defenses. In addition, effective attack strategies can be made by adjusting attack data distribution. Finally, we discuss the potential directions of defending the curses brought by data heterogeneity. The results and lessons learned from our extensive experiments and analysis offer new insights for designing robust federated learning methods and systems.
23
Gonzalez-Cuautle, David, Aldo Hernandez-Suarez, Gabriel Sanchez-Perez, Linda Karina Toscano-Medina, Jose Portillo-Portillo, Jesus Olivares-Mercado, Hector Manuel Perez-Meana e Ana Lucila Sandoval-Orozco. "Synthetic Minority Oversampling Technique for Optimizing Classification Tasks in Botnet and Intrusion-Detection-System Datasets". Applied Sciences 10, n. 3 (22 gennaio 2020): 794. http://dx.doi.org/10.3390/app10030794.
Abstract (sommario):
Presently, security is a hot research topic due to the impact in daily information infrastructure. Machine-learning solutions have been improving classical detection practices, but detection tasks employ irregular amounts of data since the number of instances that represent one or several malicious samples can significantly vary. In highly unbalanced data, classification models regularly have high precision with respect to the majority class, while minority classes are considered noise due to the lack of information that they provide. Well-known datasets used for malware-based analyses like botnet attacks and Intrusion Detection Systems (IDS) mainly comprise logs, records, or network-traffic captures that do not provide an ideal source of evidence as a result of obtaining raw data. As an example, the numbers of abnormal and constant connections generated by either botnets or intruders within a network are considerably smaller than those from benign applications. In most cases, inadequate dataset design may lead to the downgrade of a learning algorithm, resulting in overfitting and poor classification rates. To address these problems, we propose a resampling method, the Synthetic Minority Oversampling Technique (SMOTE) with a grid-search algorithm optimization procedure. This work demonstrates classification-result improvements for botnet and IDS datasets by merging synthetically generated balanced data and tuning different supervised-learning algorithms.
24
Salama, Wessam M., Moustafa H. Aly e Azza M. Elbagoury. "Lung Images Segmentation and Classification Based on Deep Learning: A New Automated CNN Approach". Journal of Physics: Conference Series 2128, n. 1 (1 dicembre 2021): 012011. http://dx.doi.org/10.1088/1742-6596/2128/1/012011.
Abstract (sommario):
Abstract Lung cancer became a significant health problem worldwide over the past decades. This paper introduces a new generalized framework for lung cancer detection where many different strategies are explored for the classification. The ResNet50 model is applied to classify CT lung images into benign or malignant. Also, the U-Net, which is one of the most used architectures in deep learning for image segmentation, is employed to segment CT images before classification to increase system performance. Moreover, Image Size Dependent Normalization Technique (ISDNT) and Wiener filter are utilized as the preprocessing phase to enhance the images and suppress the noise. Our proposed framework which comprises preprocessing, segmentation and classification phases, is applied on two databases: Lung Nodule Analysis 2016 (Luna 16) and National Lung Screening Trial (NLST). Data augmentation technique is applied to solve the problem of lung CT images deficiency, and consequently, the overfitting of deep models will be avoided. The classification results show that the preprocessing for the CT lung image as the input for ResNet50-U-Net hybrid model achieves the best performance. The proposed model achieves 98.98% accuracy (ACC), 98.65% area under the ROC curve (AUC), 98.99% sensitivity (Se), 98.43% precision (Pr), 98.86% F1- score and 1.9876 s computational time.
25
Balasubramaniam, Sathiyabhama, Yuvarajan Velmurugan, Dhayanithi Jaganathan e Seshathiri Dhanasekaran. "A Modified LeNet CNN for Breast Cancer Diagnosis in Ultrasound Images". Diagnostics 13, n. 17 (24 agosto 2023): 2746. http://dx.doi.org/10.3390/diagnostics13172746.
Abstract (sommario):
Convolutional neural networks (CNNs) have been extensively utilized in medical image processing to automatically extract meaningful features and classify various medical conditions, enabling faster and more accurate diagnoses. In this paper, LeNet, a classic CNN architecture, has been successfully applied to breast cancer data analysis. It demonstrates its ability to extract discriminative features and classify malignant and benign tumors with high accuracy, thereby supporting early detection and diagnosis of breast cancer. LeNet with corrected Rectified Linear Unit (ReLU), a modification of the traditional ReLU activation function, has been found to improve the performance of LeNet in breast cancer data analysis tasks via addressing the “dying ReLU” problem and enhancing the discriminative power of the extracted features. This has led to more accurate, reliable breast cancer detection and diagnosis and improved patient outcomes. Batch normalization improves the performance and training stability of small and shallow CNN architecture like LeNet. It helps to mitigate the effects of internal covariate shift, which refers to the change in the distribution of network activations during training. This classifier will lessen the overfitting problem and reduce the running time. The designed classifier is evaluated against the benchmarking deep learning models, proving that this has produced a higher recognition rate. The accuracy of the breast image recognition rate is 89.91%. This model will achieve better performance in segmentation, feature extraction, classification, and breast cancer tumor detection.
26
Radhi, Eman, e Mohammed Kamil. "An automatic segmentation of breast ultrasound images using U-Net model". Serbian Journal of Electrical Engineering 20, n. 2 (2023): 191–203. http://dx.doi.org/10.2298/sjee2302191r.
Abstract (sommario):
Medical imaging, like ultrasound, gives a good visual picture of how an organ works. However, a radiologist has a hard time and takes a long time to process these images, which delays the diagnosis. Several automated methods for detecting and segmenting breast lesions have been developed. Nevertheless, due to ultrasonic artifacts and the intricacy of lesion forms and locations, the segmentation of lesions or tumors from breast ultrasonography remains an open issue. Medical image segmentation has seen a breakthrough thanks to deep learning. U-Net is the most noteworthy deep network in this regard. The traditional U-Net design lacks precision when dealing with complex data sets, despite its exceptional performance in segmenting multimedia medical images. To reduce texture detail redundancy and avoid overfitting, we suggest developing the U-Net architecture by including dropout layers after each max pooling layer. Batchnormalization layers and a binary cross-entropy loss function were used to preserve breast tumor texture features and edge attributes while decreasing computational costs. We used the breast ultrasound dataset of 780 images with normal, benign, or malignant tumors. Our model showed superior segmentation results for breast ultrasound pictures compared to previous deep neural networks. Quantitative measures, accuracy, and IoU values were utilized to evaluate the suggested model?s effectiveness. The results were 99.34% and 99.60% for accuracy and IoU. The results imply that the augmented U-Net model that has been suggested has high diagnostic potential in the clinic since it can correctly segment breast lesions.
27
Kujdowicz, Monika, Dominika Januś, Anna Taczanowska-Niemczuk, Marek W. Lankosz e Dariusz Adamek. "Raman Spectroscopy as a Potential Adjunct of Thyroid Nodule Evaluation: A Systematic Review". International Journal of Molecular Sciences 24, n. 20 (13 ottobre 2023): 15131. http://dx.doi.org/10.3390/ijms242015131.
Abstract (sommario):
The incidence of thyroid nodules (TNs) is estimated at 36.5% and 23% in females and males, respectively. A single thyroid nodule is usually detected during ultrasound assessment in patients with symptoms of thyroid dysfunction or neck mass. TNs are classified as benign tumours (non-malignant hyperplasia), benign neoplasms (e.g., adenoma, a non-invasive follicular tumour with papillary nuclear features) or malignant carcinomas (follicular cell-derived or C-cell derived). The differential diagnosis is based on fine-needle aspiration biopsies and cytological assessment (which is burdened with the bias of subjectivity). Raman spectroscopy (RS) is a laser-based, semiquantitative technique which shows for oscillations of many chemical groups in one label-free measurement. RS, through the assessment of chemical content, gives insight into tissue state which, in turn, allows for the differentiation of disease on the basis of spectral characteristics. The purpose of this study was to report if RS could be useful in the differential diagnosis of TN. The Web of Science, PubMed, and Scopus were searched from the beginning of the databases up to the end of June 2023. Two investigators independently screened key data using the terms “Raman spectroscopy” and “thyroid”. From the 4046 records found initially, we identified 19 studies addressing the differential diagnosis of TNs applying the RS technique. The lasers used included 532, 633, 785, 830, and 1064 nm lines. The thyroid RS investigations were performed at the cellular and/or tissue level, as well as in serum samples. The accuracy of papillary thyroid carcinoma detection is approx. 90%. Furthermore, medullary, and follicular thyroid carcinoma can be detected with up to 100% accuracy. These results might be biased with low numbers of cases in some research and overfitting of models as well as the reference method. The main biochemical changes one can observe in malignancies are as follows: increase of protein, amino acids (like phenylalanine, tyrosine, and tryptophan), and nucleic acid content in comparison with non-malignant TNs. Herein, we present a review of the literature on the application of RS in the differential diagnosis of TNs. This technique seems to have powerful application potential in thyroid tumour diagnosis.
28
Alhussainan, Norah Fahd, Belgacem Ben Youssef e Mohamed Maher Ben Ismail. "A Deep Learning Approach for Brain Tumor Firmness Detection Based on Five Different YOLO Versions: YOLOv3–YOLOv7". Computation 12, n. 3 (1 marzo 2024): 44. http://dx.doi.org/10.3390/computation12030044.
Abstract (sommario):
Brain tumor diagnosis traditionally relies on the manual examination of magnetic resonance images (MRIs), a process that is prone to human error and is also time consuming. Recent advancements leverage machine learning models to categorize tumors, such as distinguishing between “malignant” and “benign” classes. This study focuses on the supervised machine learning task of classifying “firm” and “soft” meningiomas, critical for determining optimal brain tumor treatment. The research aims to enhance meningioma firmness detection using state-of-the-art deep learning architectures. The study employs a YOLO architecture adapted for meningioma classification (Firm vs. Soft). This YOLO-based model serves as a machine learning component within a proposed CAD system. To improve model generalization and combat overfitting, transfer learning and data augmentation techniques are explored. Intra-model analysis is conducted for each of the five YOLO versions, optimizing parameters such as the optimizer, batch size, and learning rate based on sensitivity and training time. YOLOv3, YOLOv4, and YOLOv7 demonstrate exceptional sensitivity, reaching 100%. Comparative analysis against state-of-the-art models highlights their superiority. YOLOv7, utilizing the SGD optimizer, a batch size of 64, and a learning rate of 0.01, achieves outstanding overall performance with metrics including mean average precision (99.96%), precision (98.50%), specificity (97.95%), balanced accuracy (98.97%), and F1-score (99.24%). This research showcases the effectiveness of YOLO architectures in meningioma firmness detection, with YOLOv7 emerging as the optimal model. The study’s findings underscore the significance of model selection and parameter optimization for achieving high sensitivity and robust overall performance in brain tumor classification.
29
Wang, Ruikui, Yuanfang Guo e Yunhong Wang. "AGS: Affordable and Generalizable Substitute Training for Transferable Adversarial Attack". Proceedings of the AAAI Conference on Artificial Intelligence 38, n. 6 (24 marzo 2024): 5553–62. http://dx.doi.org/10.1609/aaai.v38i6.28365.
Abstract (sommario):
In practical black-box attack scenarios, most of the existing transfer-based attacks employ pretrained models (e.g. ResNet50) as the substitute models. Unfortunately, these substitute models are not always appropriate for transfer-based attacks. Firstly, these models are usually trained on a largescale annotated dataset, which is extremely expensive and time-consuming to construct. Secondly, the primary goal of these models is to perform a specific task, such as image classification, which is not developed for adversarial attacks. To tackle the above issues, i.e., high cost and over-fitting on taskspecific models, we propose an Affordable and Generalizable Substitute (AGS) training framework tailored for transferbased adversarial attack. Specifically, we train the substitute model from scratch by our proposed adversary-centric constrastive learning. This proposed learning mechanism introduces another sample with slight adversarial perturbations as an additional positive view of the input image, and then encourages the adversarial view and two benign views to interact comprehensively with each other. To further boost the generalizability of the substitute model, we propose adversarial invariant learning to maintain the representations of the adversarial example invariants under augmentations with various strengths. Our AGS model can be trained solely with unlabeled and out-of domain data and avoid overfitting to any task-specific models, because of its inherently self-supervised nature. Extensive experiments demonstrate that our AGS achieves comparable or superior performance compared to substitute models pretrained on the complete ImageNet training set, when executing attacks across a diverse range of target models, including ViTs, robustly trained models, object detection and segmentation models. Our source codes are available at https://github.com/lwmming/AGS.
30
Shah, Rajesh P., Heather M. Selby, Pritam Mukherjee, Shefali Verma, Peiyi Xie, Qinmei Xu, Millie Das, Sachin Malik, Olivier Gevaert e Sandy Napel. "Machine Learning Radiomics Model for Early Identification of Small-Cell Lung Cancer on Computed Tomography Scans". JCO Clinical Cancer Informatics, n. 5 (giugno 2021): 746–57. http://dx.doi.org/10.1200/cci.21.00021.
Abstract (sommario):
PURPOSE Small-cell lung cancer (SCLC) is the deadliest form of lung cancer, partly because of its short doubling time. Delays in imaging identification and diagnosis of nodules create a risk for stage migration. The purpose of our study was to determine if a machine learning radiomics model can detect SCLC on computed tomography (CT) among all nodules at least 1 cm in size. MATERIALS AND METHODS Computed tomography scans from a single institution were selected and resampled to 1 × 1 × 1 mm. Studies were divided into SCLC and other scans comprising benign, adenocarcinoma, and squamous cell carcinoma that were segregated into group A (noncontrast scans) and group B (contrast-enhanced scans). Four machine learning classification models, support vector classifier, random forest (RF), XGBoost, and logistic regression, were used to generate radiomic models using 59 quantitative first-order and texture Imaging Biomarker Standardization Initiative compliant PyRadiomics features, which were found to be robust between two segmenters with minimum Redundancy Maximum Relevance feature selection within each leave-one-out-cross-validation to avoid overfitting. The performance was evaluated using a receiver operating characteristic curve. A final model was created using the RF classifier and aggregate minimum Redundancy Maximum Relevance to determine feature importance. RESULTS A total of 103 studies were included in the analysis. The area under the receiver operating characteristic curve for RF, support vector classifier, XGBoost, and logistic regression was 0.81, 0.77, 0.84, and 0.84 in group A, and 0.88, 0.87, 0.85, and 0.81 in group B, respectively. Nine radiomic features in group A and 14 radiomic features in group B were predictive of SCLC. Six radiomic features overlapped between groups A and B. CONCLUSION A machine learning radiomics model may help differentiate SCLC from other lung lesions.
31
Alzubaidi, Laith, Omran Al-Shamma, Mohammed A. Fadhel, Laith Farhan, Jinglan Zhang e Ye Duan. "Optimizing the Performance of Breast Cancer Classification by Employing the Same Domain Transfer Learning from Hybrid Deep Convolutional Neural Network Model". Electronics 9, n. 3 (6 marzo 2020): 445. http://dx.doi.org/10.3390/electronics9030445.
Abstract (sommario):
Breast cancer is a significant factor in female mortality. An early cancer diagnosis leads to a reduction in the breast cancer death rate. With the help of a computer-aided diagnosis system, the efficiency increased, and the cost was reduced for the cancer diagnosis. Traditional breast cancer classification techniques are based on handcrafted features techniques, and their performance relies upon the chosen features. They also are very sensitive to different sizes and complex shapes. However, histopathological breast cancer images are very complex in shape. Currently, deep learning models have become an alternative solution for diagnosis, and have overcome the drawbacks of classical classification techniques. Although deep learning has performed well in various tasks of computer vision and pattern recognition, it still has some challenges. One of the main challenges is the lack of training data. To address this challenge and optimize the performance, we have utilized a transfer learning technique which is where the deep learning models train on a task, and then fine-tune the models for another task. We have employed transfer learning in two ways: Training our proposed model first on the same domain dataset, then on the target dataset, and training our model on a different domain dataset, then on the target dataset. We have empirically proven that the same domain transfer learning optimized the performance. Our hybrid model of parallel convolutional layers and residual links is utilized to classify hematoxylin–eosin-stained breast biopsy images into four classes: invasive carcinoma, in-situ carcinoma, benign tumor and normal tissue. To reduce the effect of overfitting, we have augmented the images with different image processing techniques. The proposed model achieved state-of-the-art performance, and it outperformed the latest methods by achieving a patch-wise classification accuracy of 90.5%, and an image-wise classification accuracy of 97.4% on the validation set. Moreover, we have achieved an image-wise classification accuracy of 96.1% on the test set of the microscopy ICIAR-2018 dataset.
32
Wildeboer, Rogier R., Christophe K. Mannaerts, Ruud J. G. van Sloun, Lars Budäus, Derya Tilki, Hessel Wijkstra, Georg Salomon e Massimo Mischi. "Automated multiparametric localization of prostate cancer based on B-mode, shear-wave elastography, and contrast-enhanced ultrasound radiomics". European Radiology 30, n. 2 (10 ottobre 2019): 806–15. http://dx.doi.org/10.1007/s00330-019-06436-w.
Abstract (sommario):
Abstract Objectives The aim of this study was to assess the potential of machine learning based on B-mode, shear-wave elastography (SWE), and dynamic contrast-enhanced ultrasound (DCE-US) radiomics for the localization of prostate cancer (PCa) lesions using transrectal ultrasound. Methods This study was approved by the institutional review board and comprised 50 men with biopsy-confirmed PCa that were referred for radical prostatectomy. Prior to surgery, patients received transrectal ultrasound (TRUS), SWE, and DCE-US for three imaging planes. The images were automatically segmented and registered. First, model-based features related to contrast perfusion and dispersion were extracted from the DCE-US videos. Subsequently, radiomics were retrieved from all modalities. Machine learning was applied through a random forest classification algorithm, using the co-registered histopathology from the radical prostatectomy specimens as a reference to draw benign and malignant regions of interest. To avoid overfitting, the performance of the multiparametric classifier was assessed through leave-one-patient-out cross-validation. Results The multiparametric classifier reached a region-wise area under the receiver operating characteristics curve (ROC-AUC) of 0.75 and 0.90 for PCa and Gleason > 3 + 4 significant PCa, respectively, thereby outperforming the best-performing single parameter (i.e., contrast velocity) yielding ROC-AUCs of 0.69 and 0.76, respectively. Machine learning revealed that combinations between perfusion-, dispersion-, and elasticity-related features were favored. Conclusions In this paper, technical feasibility of multiparametric machine learning to improve upon single US modalities for the localization of PCa has been demonstrated. Extended datasets for training and testing may establish the clinical value of automatic multiparametric US classification in the early diagnosis of PCa. Key Points • Combination of B-mode ultrasound, shear-wave elastography, and contrast ultrasound radiomics through machine learning is technically feasible. • Multiparametric ultrasound demonstrated a higher prostate cancer localization ability than single ultrasound modalities. • Computer-aided multiparametric ultrasound could help clinicians in biopsy targeting.
33
Atarsaikhan, Gantugs, Isabel Mogollon, Katja Välimäki, Tuomas Mirtti, Teijo Pellinen e Lassi Paavolainen. "Abstract 892: Pan-cancer tumor microenvironment profiling with multiplexed immunofluorescence microscopy and self-supervised learning". Cancer Research 84, n. 6_Supplement (22 marzo 2024): 892. http://dx.doi.org/10.1158/1538-7445.am2024-892.
Abstract (sommario):
Abstract Multiplexed immunofluorescence (mIF) microscopy reveals the spatial architecture of cancer tissue and its microenvironment that is not being fully explored with existing analysis methods. Most analysis approaches for mIF microscopy data focus on single-cell or region classification and rely on supervised machine learning. In this work, we present a self-supervised spatial profiling method for mIF images that takes local and global associations into account, and apply this method to profile cancer-associated fibroblasts (CAFs) in a pan-cancer dataset. We studied a 2-stage self-supervised training scheme to learn the representations of mIF tissue microarray (TMA) images in local and global scales (cellular and long-range associations). During the 1st self-supervised training stage, a Vision Transformer learns the local-scale representations by using small patches from TMA images. Then, the local representations are used as input to the 2nd stage where a similar self-supervised learning strategy is used to learn global patterns in the TMA images. We applied the method to profile multiple cohorts from three different solid tumors: prostate, renal, and lung cancer. In total, these cohorts include more than 5,000 TMA cores from over 1,750 patients extracted from the tumor center, tumor edge, and adjacent benign areas. The samples were stained with a CAF panel including FAP, aSMA, PDGFRB, pSTAT3/PDGFRA, nuclear and epithelial markers, and imaged with cyclic mIF microscopy. Samples were studied at the patch-level and TMA core-level. Small patches enable further analysis of associations in the local environment, whereas the core-level enables associations with patient clinical information. Clustering of patch-level (1st stage) and core-level (2nd stage) representations showed independently that self-supervised learning is capable of learning the representations of the mIF images. We were able to identify regions and cases with high pTNM staging from the prostate cancer samples and similar histological subtypes from renal cancer samples. We further validated the clustering using k-NN classification that showed high classification accuracy in all cohorts. Moreover, we developed a stopping criteria for the final model selection that balances between the similarities of samples and patches inside of samples to prevent overfitting. Our study shows that self-supervised learning enables unbiased discoveries from large-scale mIF microscopy imaging datasets. The developed method uncovers associations between imaging data and clinical information and highlights directly the patterns that are most meaningful for these associations. Citation Format: Gantugs Atarsaikhan, Isabel Mogollon, Katja Välimäki, Tuomas Mirtti, Teijo Pellinen, Lassi Paavolainen. Pan-cancer tumor microenvironment profiling with multiplexed immunofluorescence microscopy and self-supervised learning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 892.
34
Feng, Liqi, Yaqin Zhao, Yichao Sun, Wenxuan Zhao e Jiaxi Tang. "Action Recognition Using a Spatial-Temporal Network for Wild Felines". Animals 11, n. 2 (12 febbraio 2021): 485. http://dx.doi.org/10.3390/ani11020485.
Abstract (sommario):
Behavior analysis of wild felines has significance for the protection of a grassland ecological environment. Compared with human action recognition, fewer researchers have focused on feline behavior analysis. This paper proposes a novel two-stream architecture that incorporates spatial and temporal networks for wild feline action recognition. The spatial portion outlines the object region extracted by Mask region-based convolutional neural network (R-CNN) and builds a Tiny Visual Geometry Group (VGG) network for static action recognition. Compared with VGG16, the Tiny VGG network can reduce the number of network parameters and avoid overfitting. The temporal part presents a novel skeleton-based action recognition model based on the bending angle fluctuation amplitude of the knee joints in a video clip. Due to its temporal features, the model can effectively distinguish between different upright actions, such as standing, ambling, and galloping, particularly when the felines are occluded by objects such as plants, fallen trees, and so on. The experimental results showed that the proposed two-stream network model can effectively outline the wild feline targets in captured images and can significantly improve the performance of wild feline action recognition due to its spatial and temporal features.
35
Tran-Quoc, Kim, Lieu B. Nguyen, Van Hai Luong e H. Nguyen-Xuan. "Machine learning for predicting mechanical behavior of concrete beams with 3D printed TPMS". Vietnam Journal of Mechanics 44, n. 4 (31 dicembre 2022): 538–84. http://dx.doi.org/10.15625/0866-7136/17999.
Abstract (sommario):
Bioinspired structures are remarkable porous structures with great strength-to-weight ratios. Hence, they have been applied in various fields including biomedical, transportation, and aerospace materials, etc. Recent studies have shown the significant impact of the plastic 3D printed triply periodic minimal surfaces (TPMS) structure on the cement beam including increasing the peak load, reducing the deflection, and improving the ductility. In this study, a machine learning (ML) surrogate model has been conducted to predict the beam behavior under static bending load. At first, various combinations of plastic volume fractions and numbers of core layers have been adopted to reinforce the constituent beam. The finite element method (FEM) was implemented to investigate the influences of these reinforcement strategies. Next, the above data were employed to create the ML model. A three-process assessment was proposed to achieve the most suitable model for the present problem, these processes were the model hyperparameter tuning, the performance assessment, and the handling overfitting with deep learning (DL) techniques. Consequently, both beam peak loads and maximum deflections were proportional to the volume fraction. The increment in TPMS layers could lead to the enhancement in both traits but with a nonlinear relationship. Furthermore, each trait may be a ceiling value that could not be exceeded with a specific volume fraction despite any number of layers. This conclusion was indicated by the surrogate model predictions. The final model in this study could deal with noisy data from FEM and with the support of a new early stopping condition, excellent performance could be found on both train and test data. The maximum deviations of 2.5% and 3.5% for peak loads and maximum midpoint displacements, respectively, have verified the robustness of the present surrogate model.
36
Wang, Ke, Vidya Muthukumar e Christos Thrampoulidis. "Benign Overfitting in Multiclass Classification: All Roads Lead to Interpolation". IEEE Transactions on Information Theory, 2023, 1. http://dx.doi.org/10.1109/tit.2023.3320098.
37
Zhou, Lijia, Frederic Koehler, Danica J. Sutherland e Nathan Srebro. "Optimistic Rates: A Unifying Theory for Interpolation Learningand Regularization in Linear Regression". ACM / IMS Journal of Data Science, 16 novembre 2023. http://dx.doi.org/10.1145/3594234.
Abstract (sommario):
We study a localized notion of uniform convergence known as an “optimistic rate” [34, 39] for linear regression with Gaussian data. Our refined analysis avoids the hidden constant and logarithmic factor in existing results, which are known to be crucial in high-dimensional settings, especially for understanding interpolation learning. As a special case, our analysis recovers the guarantee from Koehler et al. [21], which tightly characterizes the population risk of low-norm interpolators under the benign overfitting conditions. Our optimistic rate bound, though, also analyzes predictors with arbitrary training error. This allows us to recover some classical statistical guarantees for ridge and LASSO regression under random designs, and helps us obtain a precise understanding of the excess risk of near-interpolators in the over-parameterized regime.
38
Zufry, Hendra, e Agus Arip Munawar. "Near-Infrared Spectroscopy for Distinguishing Malignancy in Thyroid Nodules". Applied Spectroscopy, 19 febbraio 2024. http://dx.doi.org/10.1177/00037028241232440.
Abstract (sommario):
Thyroid nodules are common clinical entities, with a significant proportion being malignant. Early, accurate, and non-invasive tools to differentiate benign and malignant nodules can optimize patient management and reduce unnecessary surgery. This study aimed to evaluate the efficacy and accuracy of near-infrared spectroscopy (NIRS) in distinguishing benign from malignant thyroid nodules. A diffuse reflectance spectrum for a total of 20 thyroid nodule samples (10 samples as colloid goiter and 10 samples as thyroid cancer), were acquired in the wavelength range from 1000 to 2500 nm. Spectral data from NIRS were analyzed by means of principal component analysis (PCA), quadratic discriminant analysis (QDA), and linear discriminant analysis (LDA) to classify and differentiate thyroid nodule samples. The present study found that NIRS effectively distinguished colloid goiter and thyroid cancer using the first two principal components (PCs), explaining 90% and 10% of the variance, respectively. QDA discrimination plot displayed a clear separation between colloid goiter and thyroid cancer with minimal overlap, aligning with reported 95% accuracy. Additionally, applying LDA to seven PCs from PCA achieved a 100% accuracy rate in classifying colloid goiter and thyroid cancer from near-infrared spectral data. In conclusion, NIRS offers a promising, non-invasive complementing diagnostic tool for differentiating benign from malignant thyroid nodules with high accuracy. Future work should integrate these results into predictive model development, emphasizing external validation, alternative performance metrics, and protecting against potential overfitting translation of a machine learning model to a clinical setting.
39
To, Tyrell, Tongtong Lu, Julie M. Jorns, Mollie Patton, Taly Gilat Schmidt, Tina Yen, Bing Yu e Dong Hye Ye. "Deep learning classification of deep ultraviolet fluorescence images toward intra-operative margin assessment in breast cancer". Frontiers in Oncology 13 (16 giugno 2023). http://dx.doi.org/10.3389/fonc.2023.1179025.
Abstract (sommario):
BackgroundBreast-conserving surgery is aimed at removing all cancerous cells while minimizing the loss of healthy tissue. To ensure a balance between complete resection of cancer and preservation of healthy tissue, it is necessary to assess themargins of the removed specimen during the operation. Deep ultraviolet (DUV) fluorescence scanning microscopy provides rapid whole-surface imaging (WSI) of resected tissues with significant contrast between malignant and normal/benign tissue. Intra-operative margin assessment with DUV images would benefit from an automated breast cancer classification method.MethodsDeep learning has shown promising results in breast cancer classification, but the limited DUV image dataset presents the challenge of overfitting to train a robust network. To overcome this challenge, the DUV-WSI images are split into small patches, and features are extracted using a pre-trained convolutional neural network—afterward, a gradient-boosting tree trains on these features for patch-level classification. An ensemble learning approach merges patch-level classification results and regional importance to determine the margin status. An explainable artificial intelligence method calculates the regional importance values.ResultsThe proposed method’s ability to determine the DUV WSI was high with 95% accuracy. The 100% sensitivity shows that the method can detect malignant cases efficiently. The method could also accurately localize areas that contain malignant or normal/benign tissue.ConclusionThe proposed method outperforms the standard deep learning classification methods on the DUV breast surgical samples. The results suggest that it can be used to improve classification performance and identify cancerous regions more effectively.
40
Kim, Taehyun, Woonyoung Chang, Jeongyoun Ahn e Sungkyu Jung. "Double data piling: a high-dimensional solution for asymptotically perfect multi-category classification". Journal of the Korean Statistical Society, 3 aprile 2024. http://dx.doi.org/10.1007/s42952-024-00263-6.
Abstract (sommario):
AbstractFor high-dimensional classification, interpolation of training data manifests as the data piling phenomenon, in which linear projections of data vectors from each class collapse to a single value. Recent research has revealed an additional phenomenon known as the ‘second data piling’ for independent test data in binary classification, providing a theoretical understanding of asymptotically perfect classification. This paper extends these findings to multi-category classification and provides a comprehensive characterization of the double data piling phenomenon. We define the maximal data piling subspace, which maximizes the sum of pairwise distances between piles of training data in multi-category classification. Furthermore, we show that a second data piling subspace that induces data piling for independent data exists and can be consistently estimated by projecting the negatively-ridged discriminant subspace onto an estimated ‘signal’ subspace. By leveraging this second data piling phenomenon, we propose a bias-correction strategy for class assignments, which asymptotically achieves perfect classification. The present research sheds light on benign overfitting and enhances the understanding of perfect multi-category classification of high-dimensional discrimination with a help of high-dimensional asymptotics.
41
Feliciani, Giacomo, Francesco Serra, Enrico Menghi, Fabio Ferroni, Anna Sarnelli, Carlo Feo, Maria Chiara Zatelli, Maria Rosaria Ambrosio, Melchiore Giganti e Aldo Carnevale. "Radiomics in the characterization of lipid-poor adrenal adenomas at unenhanced CT: time to look beyond usual density metrics". European Radiology, 11 agosto 2023. http://dx.doi.org/10.1007/s00330-023-10090-8.
Abstract (sommario):
Abstract Objectives In this study, we developed a radiomic signature for the classification of benign lipid-poor adenomas, which may potentially help clinicians limit the number of unnecessary investigations in clinical practice. Indeterminate adrenal lesions of benign and malignant nature may exhibit different values of key radiomics features. Methods Patients who had available histopathology reports and a non-contrast-enhanced CT scan were included in the study. Radiomics feature extraction was done after the adrenal lesions were contoured. The primary feature selection and prediction performance scores were calculated using the least absolute shrinkage and selection operator (LASSO). To eliminate redundancy, the best-performing features were further examined using the Pearson correlation coefficient, and new predictive models were created. Results This investigation covered 50 lesions in 48 patients. After LASSO-based radiomics feature selection, the test dataset’s 30 iterations of logistic regression models produced an average performance of 0.72. The model with the best performance, made up of 13 radiomics features, had an AUC of 0.99 in the training phase and 1.00 in the test phase. The number of features was lowered to 5 after performing Pearson’s correlation to prevent overfitting. The final radiomic signature trained a number of machine learning classifiers, with an average AUC of 0.93. Conclusions Including more radiomics features in the identification of adenomas may improve the accuracy of NECT and reduce the need for additional imaging procedures and clinical workup, according to this and other recent radiomics studies that have clear points of contact with current clinical practice. Clinical relevance statement The study developed a radiomic signature using unenhanced CT scans for classifying lipid-poor adenomas, potentially reducing unnecessary investigations that scored a final accuracy of 93%. Key Points • Radiomics has potential for differentiating lipid-poor adenomas and avoiding unnecessary further investigations. • Quadratic mean, strength, maximum 3D diameter, volume density, and area density are promising predictors for adenomas. • Radiomics models reach high performance with average AUC of 0.95 in the training phase and 0.72 in the test phase.
42
Giraldo‐Roldan, Daniela, Erin Crespo Cordeiro Ribeiro, Anna Luiza Damaceno Araújo, Paulo Victor Mendes Penafort, Viviane Mariano da Silva, Jeconias Câmara, Hélder Antônio Rebelo Pontes et al. "Deep learning applied to the histopathological diagnosis of ameloblastomas and ameloblastic carcinomas". Journal of Oral Pathology & Medicine, 15 settembre 2023. http://dx.doi.org/10.1111/jop.13481.
Abstract (sommario):
AbstractBackgroundOdontogenic tumors (OT) are composed of heterogeneous lesions, which can be benign or malignant, with different behavior and histology. Within this classification, ameloblastoma and ameloblastic carcinoma (AC) represent a diagnostic challenge in daily histopathological practice due to their similar characteristics and the limitations that incisional biopsies represent. From these premises, we wanted to test the usefulness of models based on artificial intelligence (AI) in the field of oral and maxillofacial pathology for differential diagnosis. The main advantages of integrating Machine Learning (ML) with microscopic and radiographic imaging is the ability to significantly reduce intra‐and inter observer variability and improve diagnostic objectivity and reproducibility.MethodsThirty Digitized slides were collected from different diagnostic centers of oral pathology in Brazil. After performing manual annotation in the region of interest, the images were segmented and fragmented into small patches. In the supervised learning methodology for image classification, three models (ResNet50, DenseNet, and VGG16) were focus of investigation to provide the probability of an image being classified as class0 (i.e., ameloblastoma) or class1 (i.e., Ameloblastic carcinoma).ResultsThe training and validation metrics did not show convergence, characterizing overfitting. However, the test results were satisfactory, with an average for ResNet50 of 0.75, 0.71, 0.84, 0.65, and 0.77 for accuracy, precision, sensitivity, specificity, and F1‐score, respectively.ConclusionsThe models demonstrated a strong potential of learning, but lack of generalization ability. The models learn fast, reaching a training accuracy of 98%. The evaluation process showed instability in validation; however, acceptable performance in the testing process, which may be due to the small data set. This first investigation opens an opportunity for expanding collaboration to incorporate more complementary data; as well as, developing and evaluating new alternative models.
43
Jiménez-Gaona, Yuliana, María José Rodríguez-Alvarez, Líder Escudero, Carlos Sandoval e Vasudevan Lakshminarayanan. "Ultrasound breast images denoising using generative adversarial networks (GANs)". Intelligent Data Analysis, 31 gennaio 2024, 1–18. http://dx.doi.org/10.3233/ida-230631.
Abstract (sommario):
INTRODUCTION: Ultrasound in conjunction with mammography imaging, plays a vital role in the early detection and diagnosis of breast cancer. However, speckle noise affects medical ultrasound images and degrades visual radiological interpretation. Speckle carries information about the interactions of the ultrasound pulse with the tissue microstructure, which generally causes several difficulties in identifying malignant and benign regions. The application of deep learning in image denoising has gained more attention in recent years. OBJECTIVES: The main objective of this work is to reduce speckle noise while preserving features and details in breast ultrasound images using GAN models. METHODS: We proposed two GANs models (Conditional GAN and Wasserstein GAN) for speckle-denoising public breast ultrasound databases: BUSI, DATASET A, AND UDIAT (DATASET B). The Conditional GAN model was trained using the Unet architecture, and the WGAN model was trained using the Resnet architecture. The image quality results in both algorithms were measured by Peak Signal to Noise Ratio (PSNR, 35–40 dB) and Structural Similarity Index (SSIM, 0.90–0.95) standard values. RESULTS: The experimental analysis clearly shows that the Conditional GAN model achieves better breast ultrasound despeckling performance over the datasets in terms of PSNR = 38.18 dB and SSIM = 0.96 with respect to the WGAN model (PSNR = 33.0068 dB and SSIM = 0.91) on the small ultrasound training datasets. CONCLUSIONS: The observed performance differences between CGAN and WGAN will help to better implement new tasks in a computer-aided detection/diagnosis (CAD) system. In future work, these data can be used as CAD input training for image classification, reducing overfitting and improving the performance and accuracy of deep convolutional algorithms.
44
Yang, Fan, Yujie Li, Xiaolu Li, Xiaoduo Yu, Yanfeng Zhao, Lin Li, Lizhi Xie e Meng Lin. "The utility of texture analysis based on quantitative synthetic magnetic resonance imaging in nasopharyngeal carcinoma: a preliminary study". BMC Medical Imaging 23, n. 1 (25 gennaio 2023). http://dx.doi.org/10.1186/s12880-023-00968-w.
Abstract (sommario):
Abstract Background Magnetic resonance imaging (MRI) is commonly used for the diagnosis of nasopharyngeal carcinoma (NPC) and occipital clivus (OC) invasion, but a proportion of lesions may be missed using non-enhanced MRI. The purpose of this study is to investigate the diagnostic performance of synthetic magnetic resonance imaging (SyMRI) in differentiating NPC from nasopharyngeal hyperplasia (NPH), as well as evaluating OC invasion. Methods Fifty-nine patients with NPC and 48 volunteers who underwent SyMRI examination were prospectively enrolled. Eighteen first-order features were extracted from VOIs (primary tumours, benign mucosa, and OC). Statistical comparisons were conducted between groups using the independent-samples t-test and the Mann–Whitney U test to select significant parameters. Multiple diagnostic models were then constructed using multivariate logistic analysis. The diagnostic performance of the models was calculated by receiver operating characteristics (ROC) curve analysis and compared using the DeLong test. Bootstrap and 5-folds cross-validation were applied to avoid overfitting. Results The T1, T2 and PD map-derived models had excellent diagnostic performance in the discrimination between NPC and NPH in volunteers, with area under the curves (AUCs) of 0.975, 0.972 and 0.986, respectively. Besides, SyMRI models also showed excellent performance in distinguishing OC invasion from non-invasion (AUC: 0.913–0.997). Notably, the T1 map-derived model showed the highest diagnostic performance with an AUC, sensitivity, specificity, and accuracy of 0.997, 96.9%, 97.9% and 97.5%, respectively. By using 5-folds cross-validation, the bias-corrected AUCs were 0.965–0.984 in discriminating NPC from NPH and 0.889–0.975 in discriminating OC invasion from OC non-invasion. Conclusions SyMRI combined with first-order parameters showed excellent performance in differentiating NPC from NPH, as well as discriminating OC invasion from non-invasion.
45
Montaha, Sidratul, Sami Azam, Md Rahad Islam Bhuiyan, Sadia Sultana Chowa, Md Saddam Hossain Mukta e Mirjam Jonkman. "Malignancy pattern analysis of breast ultrasound images using clinical features and a graph convolutional network". DIGITAL HEALTH 10 (gennaio 2024). http://dx.doi.org/10.1177/20552076241251660.
Abstract (sommario):
Objective Early diagnosis of breast cancer can lead to effective treatment, possibly increase long-term survival rates, and improve quality of life. The objective of this study is to present an automated analysis and classification system for breast cancer using clinical markers such as tumor shape, orientation, margin, and surrounding tissue. The novelty and uniqueness of the study lie in the approach of considering medical features based on the diagnosis of radiologists. Methods Using clinical markers, a graph is generated where each feature is represented by a node, and the connection between them is represented by an edge which is derived through Pearson's correlation method. A graph convolutional network (GCN) model is proposed to classify breast tumors into benign and malignant, using the graph data. Several statistical tests are performed to assess the importance of the proposed features. The performance of the proposed GCN model is improved by experimenting with different layer configurations and hyper-parameter settings. Results Results show that the proposed model has a 98.73% test accuracy. The performance of the model is compared with a graph attention network, a one-dimensional convolutional neural network, and five transfer learning models, ten machine learning models, and three ensemble learning models. The performance of the model was further assessed with three supplementary breast cancer ultrasound image datasets, where the accuracies are 91.03%, 94.37%, and 89.62% for Dataset A, Dataset B, and Dataset C (combining Dataset A and Dataset B) respectively. Overfitting issues are assessed through k-fold cross-validation. Conclusion Several variants are utilized to present a more rigorous and fair evaluation of our work, especially the importance of extracting clinically relevant features. Moreover, a GCN model using graph data can be a promising solution for an automated feature-based breast image classification system.
46
Singhal, Aneesh B., Oguzhan Kursun, Mehmet A. Topcuoglu, Joshua Fok, Bruce Barton e Susanne Muehlschlegel. "Abstract WP431: Distinguishing RCVS-associated Subarachnoid Hemorrhage From Cryptogenic and Aneurysmal Subarachnoid Hemorrhage". Stroke 44, suppl_1 (febbraio 2013). http://dx.doi.org/10.1161/str.44.suppl_1.awp431.
Abstract (sommario):
Background: Reversible cerebral vasoconstriction syndrome (RCVS) is a self-limited entity with usually benign outcome. Over 30% RCVS patients develop subarachnoid hemorrhage (SAH). We aimed to identify features that differentiate RCVS-SAH from more ominous causes of SAH, i.e. aneurysmal SAH (aSAH) and cryptogenic ‘angio-negative’ SAH (cSAH). Methods: We compared the clinical-imaging features of 38 consecutive RCVS-SAH patients, to 515 aSAH and 93 cSAH patients consecutively admitted to Massachusetts General Hospital. Results: As compared to aSAH and cSAH, the RCVS-SAH group was significantly younger, more women, and higher frequency of migraine, depression, chronic obstructive pulmonary disease (COPD), alcohol and drug exposure, and prior antidepressant use. The distribution of Hunt-Hess (HH) grade and Fisher group were different between groups, with median values highest in the aSAH group. The RCVS-SAH group had more hypodense lesions on 1st head CT and earlier, more severe and widespread vasoconstriction on cerebral angiography. Discharge mRS scores were lowest in the RCVS-SAH group. To avoid overfitting, multivariate logistic regression (Firth’s method) was performed using separate models for clinical and radiological variables given small "N". Predictors of RCVS-SAH vs. aSAH [model 1]: age (O.R. 0.9, 95% C.I. 0.9-0.96), prior headache disorder (O.R. 9.3, 95% C.I. 3.9-22.4), depression (O.R. 5.6, 95% C.I. 1.8-17.6), and COPD (O.R. 7.6, 95% C.I. 2.9-20.1), and [model 2]: HH grade (O.R. 0.4, 95% C.I. 0.2-0.7), Fisher group (O.R. 0.2, 95% C.I. 0.07-0.4), and the number of constricted arteries (O.R. 1.6, 95% C.I. 1.4-1.9). Predictors of RCVS-SAH vs. cSAH [model 1]: age (O.R. 0.9, 95%C.I. 0.9-0.97), prior headache (O.R. 10.3, 95% C.I. 4.3-24.9), depression (O.R. 6.9, 95% C.I. 2.1-22.4), and alcohol use (O.R. 5.1, 95% C.I. 2.0-12.9), and [model 2]: Fisher group (O.R. 0.01, 95% C.I. 0.0-0.6), vasospasm severity (O.R. 9.1, 95% C.I. 1.4-57.2), and the number of constricted arteries (O.R. 2.0, 95% C.I. 1.2-3.1). Conclusion: Several clinical-imaging features distinguish RCVS-SAH from aSAH and cSAH. These data should prove useful to improve the diagnostic accuracy, management, and resource utilization in patients with SAH.