Section: Overall Objectives

Rendering, inpainting and super-resolution

This research axis addresses the problem of high quality reconstruction of various types of visual data after decoding. Depending on the application and the corresponding type of content (2D, 3D), various issues are being addressed. For example, to be able to render 3D scenes, depth information is associated with each view as a depth map, and transmitted in order to perform virtual view generation. Given one view with its depth information, depth image-based rendering techniques have the ability to render views in any other spatial positions. However, the issue of intermediate view reconstruction remains a difficult ill-posed problem. Most errors in the view synthesis are caused by incorrect geometry information, inaccurate camera parameters, and occlusions/disocclusions. Efficient inpainting techniques are necessary to restore disocclusions areas. Inpainting techniques are also required in transmission scenarios, where packet losses result in missing data in the video after decoding. The design of efficient mono-view and multi-view super-resolution methods is also part of the project-team objectives to improve the rendering quality, as well as to trade-off quality against transmission rate.