Journal articles on the topic 'Automatic region tagging'

Abstract. It’s easily to obtain the geometric information of terrain features in a timely manner using advanced surveying and mapping methods, but it is impossible to obtain their semantic information with low latency due to the rapid development of cities. The popularity of GPS-enabled devices and technologies provide us a large number of personal location information. Moreover, it is possible to extract the personal or group behavior pattern due to the regularity of human behavior. Those conditions make it possible to extract and identify human behavior patterns from their trajectory data. In this paper, we present an automatic semantic map generation method that extract semantic patterns and take advantage of them to tagging spatial objects in an unknown region based on known semantic patterns. We study the regularity of trajectory data and build the semantic pattern based on the regularity of human behavior. Most importantly, we use known semantic patterns to identify the semantics of the stay points in the unknown region, and use this method to realize the semantic recognition of the stay points. Results of the experiments show the effectiveness of our proposed method.

The paper presents the current version of the Slovene corpus of netspeak Janes which contains tweets, forum posts, news comments, blogs and blog comments, and user and talk pages from Wikipedia. First, we describe the harvesting procedure for each data source and provide a quantitative analysis of the corpus. Next, we present automatic and manual procedures for enriching the corpus with metadata, such as user type, gender and region, and text sentiment and standardness level. Finally, we give a detailed account of the linguistic annotation workflow which includes tokenization, sentence segmentation, rediacritisation, normalization, morphosyntactic tagging and lemmatization.

Photo-identification is an increasingly used method for the study of animal populations. Natural marks such as coloration or scale pattern to identify individuals provide an inexpensive and less invasive alternative to conventional tagging methods. Photo-identification has previously been used to distinguish individual snakes, usually by comparing the pileus region. Nevertheless, this method is seldom used in capture-recapture studies. We show the effectiveness of photo-identification in snakes using specific software for individual recognition applied to a wildlife control study of horseshoe whip snakes. Photos were analysed with Automatic Photo Identification Suite (APHIS), which allowed us to compare the variability of head scale patterns surrounding the parietal shields instead of the traditional method of using large scale groups of the pileus. APHIS correctly identified 100 % of recaptures of snakes. Although further studies are needed, the variability of the surrounding scales of the pileus region seems a robust method to identify and differentiate individuals.

In recent social media applications, descriptive information is collected through user tagging, such as face recognition, and automatic environment sensing, such as GPS. There are many applications that recognize landmarks using information gathered from GPS data. However, GPS is dependent on the location of the camera, not the landmark. In this research, we propose an automatic landmark tagging scheme using secondary regions to distinguish between similar landmarks. We propose two algorithms: 1) landmark tagging by secondary objects and 2) automatic new landmark recognition. Images of 30 famous landmarks from various public databases were used in our experiment. Results show increments of tagged areas and the improvement of landmark tagging accuracy.

In recent social media applications, descriptive information is collected through user tagging, such as face recognition, and automatic environment sensing, such as GPS. There are many applications that recognize landmarks using information gathered from GPS data. However, GPS is dependent on the location of the camera, not the landmark. In this research, we propose an automatic landmark tagging scheme using secondary regions to distinguish between similar landmarks. We propose two algorithms: 1) landmark tagging by secondary objects and 2) automatic new landmark recognition. Images of 30 famous landmarks from various public databases were used in our experiment. Results show increments of tagged areas and the improvement of landmark tagging accuracy.

Since rapid growth of Internet technologies and mobile devices, multimedia data such as images and videos are explosively growing on the Internet. Managing large scale multimedia data with correct tags and annotations is very important task. Incorrect tags and annotations make it hard to manage multimedia data. Accurate tags and annotation ease management of multimedia data and give high quality retrieve results. Fully manual image tagging which is tagged by user will be most accurate tags when the user tags correct information. Nevertheless, most of users do not make effort on task of tagging. Therefore, we suffer from lots of noisy tags. Best solution for accurate image tagging is to tag image automatically. Robust automatic image tagging models are proposed by many researchers and it is still most interesting research field these days. Since there are still lots of limitations in automatic image tagging models, we propose efficient automatic image tagging model using multigrid based image segmentation and feature extraction method. Our model can improve the object descriptions of images and image regions. Our method is tested with Corel dataset and the result showed that our model performance is efficient and effective compared to other models.

One of the first steps in understanding a protein's function is to determine its localization; however, the methods for localizing proteins in some systems have not kept pace with the developments in other fields, creating a bottleneck in the analysis of the large datasets that are generated in the post-genomic era. To address this, we developed tools for tagging proteins in trypanosomatids. We made a plasmid that, when coupled with long primer PCR, can be used to produce transgenes at their endogenous loci encoding proteins tagged at either terminus or within the protein coding sequence. This system can also be used to generate deletion mutants to investigate the function of different protein domains. We show that the length of homology required for successful integration precluded long primer PCR tagging in Leishmania mexicana . Hence, we developed plasmids and a fusion PCR approach to create gene tagging amplicons with sufficiently long homologous regions for targeted integration, suitable for use in trypanosomatids with less efficient homologous recombination than Trypanosoma brucei . Importantly, we have automated the primer design, developed universal PCR conditions and optimized the workflow to make this system reliable, efficient and scalable such that whole genome tagging is now an achievable goal.

This paper introduces the Electronic Repository of Greater Poland Oaths, eROThA (1386–1446), a digitisation project of a diplomatic edition of mediaeval land court oaths recorded in Latin and Old Polish, resulting in a small, lightly tagged specialised bilingual corpus. We present the background, aims, design and methodology of the project. We also discuss the problems and limitations entrenched in turning a printed diplomatic edition into a machine-readable diplomatic edition equipped with a new interpretative layer that is sensitive to the switches between Latin and Old Polish. In addition to the automatic annotation of code-switched items on the basis of typographic characteristics of the printed edition, flexible coding of recurrent language and discourse boundary phenomena has been introduced manually to account for linguistically ambiguous or neutral forms. The project offers a fully multilingual corpus, as well as customised Polish-only and Latin-only datasets, and enables filtered metadata searches in the online front-end. Overall, the report presents a methodology for constructing multilingual corpora in the context of legal cultures in medieval Central Europe that may be extrapolated to datasets originating in other periods and regions.

Cervical cancer is a worldwide public health problem with a high rate of illness and mortality among women. In this study, we proposed a novel framework based on Faster RCNN-FPN architecture for the detection of abnormal cervical cells in cytology images from a cancer screening test. We extended the Faster RCNN-FPN model by infusing deformable convolution layers into the feature pyramid network (FPN) to improve scalability. Furthermore, we introduced a global contextual aware module alongside the Region Proposal Network (RPN) to enhance the spatial correlation between the background and the foreground. Extensive experimentations with the proposed deformable and global context aware (DGCA) RCNN were carried out using the cervical image dataset of “Digital Human Body” Vision Challenge from the Alibaba Cloud TianChi Company. Performance evaluation based on the mean average precision (mAP) and receiver operating characteristic (ROC) curve has demonstrated considerable advantages of the proposed framework. Particularly, when combined with tagging of the negative image samples using traditional computer-vision techniques, 6–9% increase in mAP has been achieved. The proposed DGCA-RCNN model has potential to become a clinically useful AI tool for automated detection of cervical cancer cells in whole slide images of Pap smear.

In this paper, we propose an original deep learning framework for the automated counting and geolocation of palm trees from aerial images using convolutional neural networks. For this purpose, we collected aerial images from two different regions in Saudi Arabia, using two DJI drones, and we built a dataset of around 11,000 instances of palm trees. Then, we applied several recent convolutional neural network models (Faster R-CNN, YOLOv3, YOLOv4, and EfficientDet) to detect palms and other trees, and we conducted a complete comparative evaluation in terms of average precision and inference speed. YOLOv4 and EfficientDet-D5 yielded the best trade-off between accuracy and speed (up to 99% mean average precision and 7.4 FPS). Furthermore, using the geotagged metadata of aerial images, we used photogrammetry concepts and distance corrections to automatically detect the geographical location of detected palm trees. This geolocation technique was tested on two different types of drones (DJI Mavic Pro and Phantom 4 pro) and was assessed to provide an average geolocation accuracy that attains 1.6 m. This GPS tagging allows us to uniquely identify palm trees and count their number from a series of drone images, while correctly dealing with the issue of image overlapping. Moreover, this innovative combination between deep learning object detection and geolocalization can be generalized to any other objects in UAV images.

The article is devoted to the analysis of the content and speech organization of the texts of accounting and reporting documents of the 18th century of the Chanceries of the Don Fund Troops of the Mikhailovsky stanitsa ataman of the State Archives of the Volgograd Region. The relevance of the research is determined by solving the problems of creating an annotated linguistic corpus of documents. Lists, catalogues, registers, inventories, statements, notes are considered from the point of view of such genre parameters as name, addressee, addressee, date and place of creation of the document. Highlighted speech markers, explicating these parameters and organizing the text. The similarities and differences of documents of different genres are revealed. It is shown that the composition of texts has a two — or three-part structure; in some documents it varies, in others it is constant; the availability of the date, the indication of the addressee also varies. Information about the enumerated objects in documents of one genre is conveyed in more or less detail, systematized or not, in text or tabular form. With the variability of texts, a tendency towards standardization of texts of one genre and differentiation of different genres was noted. The considered parameters allow to establish the genre of documents without self-naming and can be used for automated meta-tagging of texts.

Corpus preparation for low-resource languages and for development of human language technology to analyze or computationally process them is a laborious task, primarily due to the unavailability of expert linguists who are native speakers of these languages and also due to the time and resources required. Bhojpuri, Magahi, and Maithili, languages of the Purvanchal region of India (in the north-eastern parts), are low-resource languages belonging to the Indo-Aryan (or Indic) family. They are closely related to Hindi, which is a relatively high-resource language, which is why we compare them with Hindi. We collected corpora for these three languages from various sources and cleaned them to the extent possible, without changing the data in them. The text belongs to different domains and genres. We calculated some basic statistical measures for these corpora at character, word, syllable, and morpheme levels. These corpora were also annotated with parts-of-speech (POS) and chunk tags. The basic statistical measures were both absolute and relative and were expected to indicate linguistic properties, such as morphological, lexical, phonological, and syntactic complexities (or richness). The results were compared with a standard Hindi corpus. For most of the measures, we tried to match the corpus size across the languages to avoid the effect of corpus size, but in some cases it turned out that using the full corpus was better, even if sizes were very different. Although the results are not very clear, we tried to draw some conclusions about the languages and the corpora. For POS tagging and chunking, the BIS tagset was used to manually annotate the data. The POS-tagged data sizes are 16,067, 14,669, and 12,310 sentences, respectively, for Bhojpuri, Magahi, and Maithili. The sizes for chunking are 9,695 and 1,954 sentences for Bhojpuri and Maithili, respectively. The inter-annotator agreement for these annotations, using Cohen’s Kappa, was 0.92, 0.64, and 0.74, respectively, for the three languages. These (annotated) corpora have been used for developing preliminary automated tools, which include POS tagger, Chunker, and Language Identifier. We have also developed the Bilingual dictionary (Purvanchal languages to Hindi) and a Synset (that can be integrated later in the Indo-WordNet) as additional resources. The main contribution of the work is the creation of basic resources for facilitating further language processing research for these languages, providing some quantitative measures about them and their similarities among themselves and with Hindi. For similarities, we use a somewhat novel measure of language similarity based on an n-gram-based language identification algorithm. An additional contribution is providing baselines for three basic NLP applications (POS tagging, chunking, and language identification) for these closely related languages.

Abstract Background BCR-ABL1 kinase domain (KD) mutations are the most common known cause of resistance to tyrosine kinase inhibitors (TKIs) in CML. Mutation analysis is critical for selection of subsequent TKI therapy after treatment failure. Low level and compound mutants (>1 KD mutation in the same molecule) may also lead to therapy failure. However, compound and multiple polyclonal mutants cannot be distinguished by conventional methods as they determine the average genotype of all molecules. Next generation sequencing (NGS) has the potential to sensitively detect these mutants, however sequencing and PCR errors confound the detection of true, low level mutants using current approaches. Indeed, we demonstrated that the reported frequency of BCR-ABL1 compound mutants may be over estimated due to PCR recombination artifacts that mimic compound mutations (Parker Blood 2014). More reliable methods are needed to appropriately assess the impact of various mutations on patient (pt) outcome. Aim To develop a clinically applicable NGS assay that can robustly distinguish BCR-ABL1 compound and polyclonal mutants. Method We have developed a novel NGS assay termed Single Molecule Consensus Sequencing (SMCS) that involves tagging individual BCR-ABL1 cDNA molecules before library amplification, enabling identification and elimination of most PCR and sequencing errors. NGS was performed on the Illumina MiSeq, 2 x 300 bp; aa 244 - 407 of the KD was examined. Reads derived from an initial BCR-ABL1 molecule are identified bioinformatically by virtue of sharing the same tag sequence. The consensus sequence of reads with the same tag is determined using automated variant calling and filtering algorithms. The consensus sequence represents the sequence of the initial BCR-ABL1 cDNA molecule (Fig A). Results To test the validity of SMCS, we examined 10 samples lacking KD mutations and 5 mock samples created by mixing compound mutant plasmids or pt samples. Examination of raw sequencing reads revealed a complex spectrum of mutants, similar to previous clinical reports. SMCS enabled bioinformatic filtering of these artifacts, largely eliminating PCR and sequencing error, and exclusively reported the compound and polyclonal mutants known to be present in the mock samples. We estimated the background error rate to be ~2x10-5 per base. The error spectrum was consistent with DNA damage causing first round PCR errors. SMCS was used to retrospectively examine samples of 46 pts (36 CP, 2 AP, 8 BP) who were resistant to ≤4 TKIs (1st and 2nd generation). 71 mutations were previously detected by Sanger sequencing in these samples, collected before starting next line TKI. Within the region examined using SMCS, there was 100% detection concordance with Sanger sequencing. We compared the results of SMCS with an amplicon NGS method performed at another centre for 24/46 pts (Ion Torrent, depth ~10000). Ion Torrent detected 34 compound mutants in 24 pts. Of the 30/34 that were within the region examined by SMCS we only detected 8. Based on observations in Parker Blood 2014, 14 of the 22 compound mutants not detected by SMCS were likely to be PCR recombination artifacts. The other 8/22 were low level (1 - 4%) and most (6/8) involved mutations rarely/never reported in TKI resistant pts so may also be artifacts (Fig B). We detected 3 additional compound mutants in these 24 pts, plus 5 in the remaining 22/46 pts. The compound mutants detected by SMCS were consistent with the pts' TKI treatment history. Conclusion We demonstrated detection of BCR-ABL1 compound and polyclonal mutants in pt samples using a novel NGS assay that has the potential to overcome technical artifacts generated with other published methods. Whilst there is no gold standard method that can accurately detect low level compound mutations, SMCS has correctly identified sequencing and PCR recombination artifacts using mock samples. The accuracy and clinical utility of SMCS for sensitive compound and polyclonal mutant detection is currently being validated in another group of 200 imatinib resistant pts. The frequency of compound mutants detected in pts with >1 mutation by SMCS in the current analysis (35%) is approximately half of that reported previously, which suggests the published frequency may have been overestimated. Our novel assay takes an important step towards enabling a more concrete understanding of the mutation spectra in pts and their association with resistance. Figure 1 Figure 1. Disclosures Yeung: Novartis: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; BMS: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding. Lustgarten:ARIAD Pharmaceuticals Inc: Employment, Equity Ownership. Hodgson:ARIAD Pharmaceuticals, Inc.: Employment, Equity Ownership. Rivera:ARIAD Pharmaceuticals Inc: Employment, Equity Ownership. Hughes:Novartis: Honoraria, Research Funding; Bristol-Myers Squibb: Honoraria, Research Funding; Ariad: Honoraria, Research Funding. Branford:Novartis: Consultancy, Honoraria, Research Funding; Bristol-Myers Squibb: Honoraria, Research Funding; Ariad: Honoraria, Research Funding; Otsuka: Honoraria, Research Funding.

Abstract Acute myeloid leukemia (AML) is a challenging disease to treat: most patients achieve remission after induction chemotherapy, but the majority eventually relapse. Minimal residual disease (MRD) after initial treatment is the best predictor of relapse and is thus a critical metric around which to develop new treatments. However, conventional MRD diagnostics, including cytology and flow cytometry, are of variable sensitivity and often only perform well in specialty centers, and there is no gold standard. Molecular tests developed to measure trace MRD in other hematological malignancies (i.e. CML) are high-resolution but assay a single, universally-present mutation, while many different genetic drivers exist in AML and these are spread among dozens of genes. Virtually every AML patient harbors a unique combination of mutations, making it difficult to design an effective universal assay. As such, most reported molecular AML MRD assays are either sensitive for mutations that are only found in a narrow subset of patients, or can screen many potential sites of mutation, but with low sensitivity. Here we present a broadly applicable Duplex Sequencing-based AML MRD assay that can readily detect mutant allele frequencies (MAF) below 1/10,000 across a large panel of genes, and below 1/100,000 in a focused panel. Conventional next generation sequencing (NGS) introduces errors during amplification and sequencing, creating a background of artifactual noise that obscures true mutations present below ~1%. Duplex Sequencing improves accuracy >100,000-fold through a molecular tagging approach whereby both strands of each original DNA duplex are ligated with a unique molecular barcode and amplified such that the reads generated from each strand can be related back to their unique original duplex. Reads can also be distinguished from those of their mate strand, thus the two strands of each DNA duplex can be compared and any discrepant nucleotide positions are discounted as errors. Our complete AML panel targets 151 exons or hotspot codons in 29 genes with a 59 kilobase (kb) hybrid-capture footprint. This region comprises loci containing single-nucleotide and short indel mutations found in approximately 90% of adult AMLs. A mean Duplex error-corrected sequence depth of 10,837 and a maximum Duplex depth of 14,967 was obtained across these targets from a single library preparation using 250 ng of sheared leukocyte DNA (Fig. 1). Duplex depth can be readily increased by preparing additional Duplex libraries from the same source DNA to achieve proportionally higher sensitivity for rarer variants. This stands in contrast to conventional NGS where, beyond a modest level, an increase in depth simply increases the number of background errors identified (Fig. 2A). We simulated low-level residual disease by mixing control DNA from a healthy young blood donor with DNA from 9 human cell lines harboring known AML mutations at dilutions from 1:100 to 1:100,000 (Table 2). The genomic loci of these 9 mutations in NRAS, KRAS and TP53 were captured with a small 1 kb probe panel. This mixture was sequenced to a mean Duplex depth of >1,000,000-fold, with the highest and lowest MAFs shown in Fig. 2B. All were close to expected frequencies (r2=0.96) with MAF as low as 6x10-6 (Fig. 3). As proof of specificity, we examined all coding nucleotide positions (excluding the 9 expected variants) and identified only 241 background variant counts out of 414,452,402 total Duplex BP, for an aggregate mutation frequency of 5.8x10-7, consistent with the estimated background of normal human aging. Our Duplex Sequencing-based AML MRD assay is flexible, broadly applicable and extremely sensitive. The assay is easily implemented using standard NGS equipment and automated cloud-based analysis software. The ~90% of AML patients served by this SNV-focused panel can be expanded to nearly 100% with complementary indel detection via targeted NGS RT-PCR. Optionally, when a patient's mutation profile from time-of-diagnosis is known, MRD testing can focus exclusively on those targets using a subset of pre-validated probes to reduce sequencing cost. Improved MRD testing will facilitate accurate prognostication, better selection among treatment options, and could serve as a surrogate endpoint in clinical trials to bring new treatments to patients faster. We are currently evaluating Duplex Sequencing MRD tests in both retrospective and prospective clinical trials. Disclosures Higgins: TwinStrand Biosciences: Employment. Williams:TwinStrand Biosciences: Employment. Buckley:CTI Biopharma: Employment; TwinStrand Biosciences: Consultancy. Radich:TwinStrand Biosciences: Research Funding. Salk:TwinStrand Biosciences: Employment, Equity Ownership.

AbstractArabic digits (1–9) are everywhere in our daily lives. These symbols convey various semantic information, and numerate adults can easily extract from them several numerical features such as magnitude and parity. Nonetheless, since most studies used active processing tasks to assess these properties, it remains unclear whether and to what degree the access to magnitude and especially to parity is automatic. Here we investigated with EEG whether spontaneous processing of magnitude or parity can be recorded in a frequency-tagging approach, in which participants are passively stimulated by fast visual sequences of Arabic digits. We assessed automatic magnitude processing by presenting a stream of frequent small digit numbers mixed with deviant large digits (and the reverse) with a sinusoidal contrast modulation at the frequency of 10 Hz. We used the same paradigm to investigate numerical parity processing, contrasting odd digits to even digits. We found significant brain responses at the frequency of the fluctuating change and its harmonics, recorded on electrodes encompassing right occipitoparietal regions, in both conditions. Our findings indicate that both magnitude and parity are spontaneously and unintentionally extracted from Arabic digits, which supports that they are salient semantic features deeply associated to digit symbols in long-term memory.

Abstract Background Mass spectrometry imaging (MSI) is a family of acquisition techniques producing images of the distribution of molecules in a sample, without any prior tagging of the molecules. This makes it a very interesting technique for exploratory research. However, the images are difficult to analyze because the enclosed data has high dimensionality, and their content does not necessarily reflect the shape of the object of interest. Conversely, magnetic resonance imaging (MRI) scans reflect the anatomy of the tissue. MRI also provides complementary information to MSI, such as the content and distribution of water. Results We propose a new workflow to merge the information from 2D MALDI–MSI and MRI images. Our workflow can be applied to large MSI datasets in a limited amount of time. Moreover, the workflow is fully automated and based on deterministic methods which ensures the reproducibility of the results. Our methods were evaluated and compared with state-of-the-art methods. Results show that the images are combined precisely and in a time-efficient manner. Conclusion Our workflow reveals molecules which co-localize with water in biological images. It can be applied on any MSI and MRI datasets which satisfy a few conditions: same regions of the shape enclosed in the images and similar intensity distributions.