M-C Larabi & S. A. Fezza

Title: Understanding binocular perception in the context of asymmetric coding and evaluation of S‐3D

Time : 10:00- 12:00  (December 9th, 2014)

Length: 2 hours

Authors: Mohamed-Chaker Larabi (chaker.larabi@univ-poitiers.fr) and Sid Ahmed Fezza

Description: Asymmetric stereoscopic coding is a very promising technique to decrease the bandwidth required for stereoscopic 3D delivery. This technique is based on the so-called “binocular suppression theory”, which specifies that if the both eyes are provided with two views having a different quality, the fused perceptual 3D image is dominated by the highest quality view. Therefore, according to this theory, one view can be encoded at a sufficiently high quality, and the other with a lower bit-budget, with transparent 3D viewing for the user. Several works have followed this concept since then.

In this tutorial, an end-to-end asymmetric stereoscopic coding system is covered. Firstly, we will give a deep description of a few keys mechanisms of binocular vision, including binocular fusion, binocular rivalry, and binocular suppression. Then we will review the state-of-the-art of asymmetric stereoscopic coding methods. We also examine the perceptual limits of asymmetric coding or the just noticeable threshold of asymmetry between views with which the stereoscopic images can be compressed. In addition, after a broad overview on the 3D image quality assessment, we emphasize on new quality assessment metrics for stereo pairs containing asymmetric distortions, which is a much more challenging problem. Finally, we will describe some consequences of prolonged exposure to asymmetric content, such as eye strain or fatigue that affect the 3D QoE of the viewer. Also, we will show the different solutions in the literature to avoid such inconveniences.

Throughout the tutorial, we will review recent research findings from psychophysical concepts to engineering models, along with a discussion of future research directions, and future outlooks for asymmetric stereoscopic coding.

Outcomes:

• An understanding of the different concepts of binocular vision.
• The state‐of‐the‐art of asymmetric stereoscopic coding methods.
• The perceptual limits of asymmetric coding or the just noticeable threshold of asymmetry.
• Quality assessment of asymmetrically distorted stereoscopic images.
• 3D QoE related to asymmetric content.

Intended Audience: This course is intended for all those with an interest in multiview coding.

Biographies:

Mohamed-Chaker Larabi (M’05-SM’07) received his PhD from the University of Poitiers (2002). He is currently the associate professor in charge of the perception, color and quality activity at the same university. His actual scientific interests deal with image and video coding and optimization, 2D and 3D image and video quality assessment, and user experience. Chaker Larabi is a member MPEG and JPEG committee (since 2000), and chair of the Advanced Image Coding group and, the Test and Quality SubGroup. He serves as a member of divisions 1 and 8 of CIE, is a member of IS&T, and a senior member of IEEE. He is involved in many local, regional, national and international projects.

Sid Ahmed Fezza (S’13) received his engineer degree from University of Dr. Tahar MOULAY (UTM), Sa”̈ıda, Algeria, in 2007, and the M.S. degree from the University of Sciences and Technologies of Oran (USTO), Oran, Algeria, in 2009, both in computer science. Since December 2009, he is PhD candidate in computer science, at Djillali Liabes University (UDL), Algeria. He is currently Assistant Professor at University of Oran, Algeria. His research interests are in the fields of image/video processing, computer vision, image and video coding, visual quality assessment, and user experience.