- All Journals
- Al-Qadisiyah University
- Basrah University
- Kirkuk University
- Mosul University
- The Iraqi Borad for Medical Specialization
- University of Anbar
- University of Technology
- Al Mustafa University
- Al-Zahraa University for Women
- Alnoor University
- Kunooz University
- Misan University
- Ninevah University
- Tikrit University
- University of Al-Hamdaniya
- University of Kerbala
Sparse Linear Prediction Based Neural Network to Conceal Missing Image Blocks | ||
| Al-Mustansiriyah Journal of Science | ||
| Article 1, Volume 27, Issue 1, 2016, Pages 96-103 | ||
| Author | ||
| Alaa Kh. Al-Azzawi | ||
| Abstract | ||
| In this paper, a linear vector predictor is exploited to limit the linear correlation between the blocks. The variance of the current vector is predicted by a feed-forward three-layered neural network. A neural network predictors based on a supervised learning paradigm was supposed to be able to exploit higher-order structural correlations between the lost block and its neighbors. Further, edge blocks can be predicted with increased accuracy. Predictors were optimized to predict horizontal, vertical, 45 , and 135 diagonally oriented edge-blocks, respectively. The proposed method achieves accurate values by estimating a predicted pixel from the available information of several past pixels of the lost block to recovering the missing coefficients. Further, the textural information in the block was used to categorize the training data. The categorizing of the blocks were based on the edge orientation in a given image block. For example, the block variances and directional variances of the four neighboring blocks. This greatly facilitated the reconstructing of the strong diagonal edges. Experimental results show that the proposed technique provides very good reconstruction capabilities for combinations of lost blocks. The perceptual quality of the predicted images is also significantly improved. In terms of loss concealment and artifacts, especially those associated with edges. | ||
|
Statistics Article View: 186 PDF Download: 25 |
||