Littérature scientifique sur le sujet « Psychovisual experiments »

Due to different visual tasks and gaze patterns, the discomfort glare experienced by pedestrians may differ from that experienced by drivers. This paper investigates the discomfort glare experienced by pedestrians under various urban LED luminaires through psychovisual experiments conducted on a test track. The ability of state-of-the-art models to predict the level of discomfort glare, measured on the de Boer rating scale, for this application is also investigated. With one exception, the models all overestimate the mean subjective discomfort glare compared to the experimental data. Models proposed by Lin et al. (2015) and Bullough et al. (2008, 2011) perform well. However, the implementation of these models is not straightforward because choices are needed to estimate some of the variables such as the background luminance and the glare source area.

A wavelet-based perceptual image coder for the compression of color images is proposed here in which the coding structure is coupled with Human Visual System models to produce high quality images. The major contribution is the development of a new model for the compression of the color components based on psychovisual experiments, which quantifies the optimum amount of compression that can be applied to the color components for a given rate. The model is developed for YCbCr color space and the perceptually uniform CIE Lab color space. A complete coding structure for the compression of color images is developed by incorporating the new perceptual model. The performance of the proposed coder is compared with a wavelet-based coder that uses the quantization stage of the JPEG2000 standard. The perceptual quality of the compressed images is tested using the wavelet-based subjective and objective perceptual quality matrices such as Mean Opinion Score, Visual Information Fidelity and Visual Signal to Noise Ratio. Though the model is developed for a perceptually lossless high quality image compression, results obtained reveal that the proposed structure gives very good perceptual quality compared to the existing schemes for lower bit rates. These advantages make the proposed coder a candidate for replacing the encoder stage of the current image compression standards.

It is substantiated that steganographic systems should be used to ensure the protection of special information resources in conditions of its prompt delivery. Here, steganographic technologies are an integral part of complex information protection systems. Simultaneously, for steganographic systems, there is a contradiction between the density of embedded data and level of information compaction of video container (level of reduction of volume bit volume of compact presented video image concerning bit volume of an initial video image). It leads to the fact that under the conditions of the required quality (reliability) of digital video information, the bit rate level of the covert channel is insufficient. Consequently, the scientific-applied problem concerns the necessity to increase the integrity (the level of correspondence of the hidden information before its embedding in a video container and after its extraction) and bit rate of the hidden channel of special information transmission. It is relevant. The solution of the described problem in the field of application of steganographic transformations can be realized based on the application of two different approaches. The first approach is based on methods of direct message embedding. But this approach is characterized by introducing distortions in the video images used as a container. Therefore, changes in structural and statistical patterns in the syntactic description of the video container happen. It reduces the potential for video container compaction. The second approach to creating steganographic transformation methods is based on information hiding using indirect embedding technique. Here, the embedding process exploits the functional dependency between the elements of the video container and the elements of the embedded message. Setting a specific dependency between the elements in the video container corresponds to the embedded element with a value of "0" or "1". However, the existing indirect steganographic transformation methods have a disadvantage. It consists of an insufficient value of embedded data density. To eliminate these disadvantages, it is proposed to develop an approach that allows using not only psychovisual but also structural redundancy of video container for concealment. Therefore, the research objective of this paper is to develop a method for indirect information withholding in the video container compression process to increase the bit rate of the hidden message channel. In the process of research, a steganographic multiagent system is constructed, which allows embedding hidden message elements without loss of information based on the indirect approach by modifying the active bases of the multiagent basis considering their uncertainty. To select transformants (data sets) as containers for information embedding, the requirement of the existence of a base system with all active bases is taken into account. The number of embedded bits of the hidden message is equal to the number of active bases in the base system of the multiadic space. Because of the made experiments, the following results have been received: in the process of embedding messages based on the created method distortions in a video container is not brought; for the created method the additional increase in the hidden channel bit rate in average 5 … 7 times are reached.

This thesis focuses on No Reference (NR) methods for Image Quality Assessment (IQA). A review of the IQA field is presented in Chapter 2; where the different IQA methods are described and classified. In particular, the application of IQA methods within a workflow chain is discussed. In Chapter 3 we focus on NR metrics for JPEG-blockiness and noise artifacts. It is in general assumed that subjective methods produce an actual estimate of the perceived quality while objective methods produce values that should be correlated with human perceptions as best as possible. From the analysis of the regression curves that correlate objective and subjective data we have found that in some cases the metric's predictions are not in correspondence with the subjective scores. After reviewing the available databases, we realize that the distortion ranges considered are not in general representative of real case applications. Therefore, in Chapter 4 the Imaging and Vision Lab (IVL) database is introduced. It was generated with the aim of assessing the quality of images corrupted by JPEG and noise. In Chapter 5 we approach the NR-IQA field by focusing on a classification problem. A framework based on machine learning classification is proposed that let us evaluate how images can be classified within different groups or classes, according to their quality. NR metrics are considered as features and the assigned classes are obtained from the psychovisual data. For the JPEG distortion case, the feature space of the classifiers is built using each NR metric as single feature and also a pool of eleven NR metrics. Classification within five and three classes was addressed. In the former case, the five classes are in correspondence to the five categories recommended by the ITU (excellent, good, fair, poor, and bad) when designing image quality experiments. In the latter case we were interested in classifying images as high, medium or low quality ones. The classifiers are trained and tested on different databases. The classifier obtained using the pool of metrics outperforms each single metric classifier. Better performance is obtained in the case of three classes. Considering an image as the combining of two signals, content and distortion, we note that the crosstalk between both signals influences both subjective and objective quality assessment. We address this problem in Chapter 6 where our working hypothesis is that regression can be improved if performed within a group of images that present similar contents in terms of low level features. The criteria chosen to divide the images in different groups is the image complexity. The proposed strategy consists on two steps: the images (of a given database) are first classified in three groups of low, medium and high complexity. In a second step, regression is performed within each of these groups separately. The strategy is tested for different NR metrics for JPEG-blockiness and noise artifacts, different databases are considered. Correlation coefficients are computed and statistical significance tests are applied. The gain in performance depends on the metric and distortion considered. Summarizing, the two main proposals of this research work, i.e. the classification approach that combines several NR metrics and the grouping strategy, are able to outperform the correlation between subjective and objective data for the case of JPEG-blockiness. Both strategies can be extended to consider other type of distortions.

Even the enormous processing capacity of the human brain is not enough to handle all the visual sensory information that falls upon the retina. Still human beings can efficiently respond to the external stimuli. Selective attention plays an important role here. It helps to select only the pertinent portions of the scene being viewed for further processing at the deeper brain. Computational modeling of this neuro-psychological phenomenon has the potential to enrich many computer vision tasks. Enormous amounts of research involving psychovisual experiments and computational models of attention have been and are being carried out all within the past few decades. This article compiles a good volume of these research efforts. It also discusses various aspects related to computational modeling of attention–such as, choice of features, evaluation of these models, and so forth.

Even the enormous processing capacity of the human brain is not enough to handle all the visual sensory information that falls upon the retina. Still human beings can efficiently respond to the external stimuli. Selective attention plays an important role here. It helps to select only the pertinent portions of the scene being viewed for further processing at the deeper brain. Computational modeling of this neuro-psychological phenomenon has the potential to enrich many computer vision tasks. Enormous amounts of research involving psychovisual experiments and computational models of attention have been and are being carried out all within the past few decades. This article compiles a good volume of these research efforts. It also discusses various aspects related to computational modeling of attention–such as, choice of features, evaluation of these models, and so forth.