Restoration of differential images for enhancement of compressed video
| Autor: | Jeongho Shin, Junghoon Jung, Shichang Joung, Joonki Paik |
|---|---|
| Rok vydání: | 2004 |
| Předmět: |
Motion compensation
business.industry ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Macroblock computer.file_format JPEG Signal Processing Media Technology Projections onto convex sets Computer vision Computer Vision and Pattern Recognition Artificial intelligence Electrical and Electronic Engineering business Quantization (image processing) computer Image restoration Data compression Mathematics Image compression |
| Zdroj: | Journal of Visual Communication and Image Representation. 15:91-109 |
| ISSN: | 1047-3203 |
| DOI: | 10.1016/j.jvcir.2003.06.002 |
| Popis: | A modified regularization algorithm is proposed to enhance compressed video by restoring predictive-coded pictures. Since most video coding standards adopt a hybrid structure of macroblock-based motion compensation and block discrete cosine transform, the blocking artifacts occurs at both the block boundary and block interior, and the degradation process due to quantization is generated on differential images. Based on this observation, a new degradation model of differential images is presented first. Then the corresponding restoration algorithm directly processes the differential images before reconstructing decoded images. Two constraints, such as directional continuities on the block boundary and on the block interior, have been used for defining convex sets for restoring differential images. The proposed differential domain restoration algorithm is compared with the corresponding reconstructed domain algorithm using the same degradation model and equivalent set of constraints. The proposed algorithm outperforms the reconstructed domain algorithm in both analytic and experimental senses. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect