Efficient Compression of Light Field
Project: Research › StUp
Compared with conventional RGB images, light field images (LFIs) contain richer scene information, which allows a wide range of interesting applications, such as post-capture refocus, change of view point, and 3-D reconstruction. However, such additional information is obtained at the cost of substantially generating more data, which poses challenges in both storage and transmission. As hand-held consumer LFI cameras have entered markets and huge amounts of light field data is being generated, compression of LF data is becoming an urgent issue. In this project, by taking the unique characteristics of LFIs into account, we plan to investigate efficient and effective methods to compress them. Specifically, we consider developing two types of computational schemes, i.e., one is a hybrid framework that effectively combines view synthesis and well-developed video encoding technologies, and the other one is an optimization-based framework, namely disparity-guided sparse low-rank matrix approximation.
|Effective start/end date||1/06/17 → …|