We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Differential pulse code modulation and motion aligned optimal reconstruction for block-based compressive video sensing using conditional autoregressive-salp swarm algorithm.
- Authors
Sekar, R.; Ravi, G.
- Abstract
Compressive Video Sensing (CVS) is the video coding framework that uses compressive sensing (CS) theory into video codec to reduce the burden of encoding. Various existing methods are developed for CVS, but enhancing reconstruction speed and reconstruction quality remains a great challenge in video compression. Hence, an optimal method is developed using the proposed Differential Pulse Code Modulation+Conditional Autoregressive-Salp Swarm (DPCM+CA-SS) model for CVS system by considering the characteristics of CVS frame. The spatial redundancy of video frames is explored by partitioning the video frames into several blocks such that each block is measured using the measurement matrix. The measurement vectors are quantized using DPCM. The key frames are compressed into bits using Huffman coding, and the packet is transmitted to the decoder side. The temporal redundancy of CVS model is explored by encoding the frames using CA-SS-based prediction model to achieve Motion Aligned (MA) optimal reconstruction. However, the proposed CA-SS is designed by the integration of Conditional Autoregressive Value at Risk (CAViaR) and Salp Swarm Algorithm (SSA), respectively. However, the proposed DPCM+CA-SS obtained maximum PSNR and SSIM of 39.6649 db and 0.9120, and the minimum total time of 4.9523 s, respectively.
- Subjects
PULSE-code modulation; VIDEO codecs; VIDEOS; HUFFMAN codes; VIDEO compression; VIDEO coding; VALUE at risk
- Publication
International Journal of Wavelets, Multiresolution & Information Processing, 2021, Vol 19, Issue 6, p1
- ISSN
0219-6913
- Publication type
Article
- DOI
10.1142/S0219691321500211