Vincent Ricordel

Quantifying how the combination of blur and disparity affects the perceived depth

The influence of a monocular depth cue, blur, on the apparent depth of stereoscopic scenes will be studied in this paper. When 3D images are shown on a planar stereoscopic display, binocular disparity becomes a pre-eminent depth cue. But it induces simultaneously the conflict between accommodation and vergence, which is often considered as a main reason for visual discomfort. If we limit this visual discomfort by decreasing the disparity, the apparent depth also decreases. We propose to decrease the (binocular) disparity of 3D presentations, and to reinforce (monocular) cues to compensate the loss of perceived depth and keep an unaltered apparent depth. We conducted a subjective experiment using a twoalternative forced choice task. Observers were required to identify the larger perceived depth in a pair of 3D images with/without blur. By fitting the result to a psychometric function, we obtained points of subjective equality in terms of disparity. We found that when blur is added to the background of the image, the viewer can perceive larger depth comparing to the images without any blur in the background. The increase of perceived depth can be considered as a function of the relative distance between the foreground and background, while it is insensitive to the distance between the viewer and the depth plane at which the blur is added.

Study of Center-Bias in the Viewing of Stereoscopic Image and a Framework for Extending 2D Visual Attention Models to 3D

Compared to the good performance that can be achieved by many 2D visual attention models, predicting salient regions of a 3D scene is still challenging. An efficient way to achieve this can be to exploit existing models designed for 2D content. However, the visual conflicts caused by binocular disparity and changes of viewing behavior in 3D viewing need to be dealt with. To cope with these, the present paper proposes a simple framework for extending 2D attention models for 3D images, well as evaluates center-bias in 3D-viewing condition. To validate the results, a database is created, which contains eye-movements of 35 subjects recorded during free viewing of eighteen 3D images and their corresponding 2D version. Fixation density maps indicate a weaker center-bias in the viewing of 3D images. Moreover, objective metric results demonstrate the efficiency of the proposed model and a large added value of center-bias when it is taken into account in computational modeling of 3D visual attention.

Spatio-temporal segmentation and regions tracking of high definition video sequences based on a Markov Random Field model

In this paper, we propose a Markov random field sequence segmentation and regions tracking model, which aims at combining color, texture, and motion features. First a motion-based segmentation is realized. Namely the global motion of the video sequence is estimated and compensated. From the remaining motion information, a rough motion segmentation is achieved. Then, we use a Markovian approach to update and track over time the video objects. The spatio-temporal map is updated and compensated using our Markov Random Field segmentation model to keep consistency in video objects tracking.

Experimenting texture similarity metric STSIM for intra prediction mode selection and block partitioning in HEVC

Textures can often be found in large areas of images and videos. They have different spectral and statistical properties as compared with normal (structural) components. Encoding them with ordinary video coders requires higher bit rate and usually results are unsatisfying in perceived quality. Recently, different perceptual tools have been developed to estimate the perceived quality of textures taking into account models of human visual system. In this paper, we investigate and discuss the practical usability of one of these tools, namely STSIM, as a distortion function for selecting the intra-prediction mode and block partitioning of texture images in HEVC. We experiment few practical implementations to examine its performance compared with default metrics used by HEVC. Experimental results showed that the perceived quality of the decoded textures has been significantly improved specially for stochastic types of textures.

Estimation of perceptual redundancies of HEVC encoded dynamic textures

Statistical redundancies have been the dominant target in the image/video compression standards. Perceptually, there exists further redundancies that can be removed to further enhance the compression efficiency. In this paper, we considered short term homogeneous patches that fall into the foveal vision as dynamic textures, for which a psychophysical test was used to estimate their amount of perceptual redundancies. We demonstrated the possible rate saving by utilizing these redundancies. We further designed a learning model that can precisely predict the amount of redundancies and accordingly proposed a generalized perceptual optimization framework.

Local texture synthesis: A static texture coding algorithm fully compatible with HEVC

Textures are one of the main characteristics of the visual scene. Perceptually, their details are less important than their semantic meaning. This property has been exploited in many texture coding (content based video coding) approaches by removing parts of the textures in the encoder and synthesizing them at the decoder side. Such an approach would necessarily need modification of the coding process and violating the standard. This paper introduces a novel algorithm for texture coding called Local Texture Synthesis (LTS), in which texture synthesis is employed in a full compatibility with HEVC standard. This implies that a basic HEVC decoder can be used to reconstruct the signal. LTS defines the necessary conditions to synthesize a patch and produces different synthesis of it. It tries then coding each of them, and finally chooses the one that minimizes the coding cost. A prototype of this algorithm, based on Markov Random Fields, is given in this paper. This prototype provides up to 10% bitrate saving (using the same quantization parameter) while maintaining an equivalent visual quality.

Vector quantization by packing of embedded truncated lattices

The purpose of this paper is to introduce a new vector quantizer (VQ) for the compression of digital image sequences. Our approach unifies both efficient coding methods: a fast lattice encoding and an unbalanced tree-structured codebook design according to a distortion vs. rate tradeoff. This tree-structured lattices VQ (TSLVQ) is based on the hierarchical packing of embedded truncated lattices. So we investigate the design of the hierarchical set of truncated lattice structures which can be optimally embedded. We present the simple quantization procedure and describe the corresponding tree-structured codebook. Finally two unbalanced tree-structured codebook design algorithms based an the BFOS distortion vs. rate criterion are used.

Modeling the perceptual distortion of dynamic textures and its application in HEVC

In the quest of perceptually optimized video coding, coding textures is representing a challenging case. While a large body of research was put into the perception of static textures, dynamic textures are still not sufficiently explored. In this paper, we focus on short term consistent patches, known as dynamic textures, with a very limited spatial and temporal extent. We estimated the visual distortion due to HEVC compression via subjective testing. The estimated distortion profile per texture was used to optimize the HEVC coding process. Experimental results showed that this technique can offer a high bitrate saving for the same subjective quality.

A foveated short term distortion model for perceptually optimized dynamic textures compression in HEVC

Due to their rapid change over time, dynamic textures represent challenging contents for the video compression standards. While several studies proposed different coding techniques, such as synthesis based coding, they mostly lack the perceptual studies on the perceived distortions. In this paper, a framework of perceptually optimized dynamic texture compression is presented. It is based on a psychophysically driven distortion model, which is utilized inside the rate distortion loop of HEVC. The model is tested for three compression levels, and showed a significant rate saving for both training and validation sequences set.

Mojette Transform on Densest Lattices in 2D and 3D

The Mojette Transform (MT) is an exact discrete form of the Radon transform. It has been originally defined on the lattice \(Z^n\) (where n is the dimension). We propose to study this transform when using the densest lattices for the dimensions 2 and 3, namely the lattice \(A^2\) and the face-centered cubic lattice \(A^3\). In order to compare the legacy MT using \(Z^n\), versus the new MT using \(A^n\), we define a fair comparison methodology between the two MT schemes. In particular we detail how to generate the projection angles by exploiting the lattice symmetries and by reordering the Haros-Farey series. Statistic criteria have been also defined to analyse the information distribution on the projections. The experimental results study shows the specific nature of the information distribution on the MT projections due to the high compacity of the \(A^n\) lattices.