Application of Artificial Neural Network in Video Compression Coding
Yong Zhang, Ming Ming Zhang · 2008
The goal of video compression is to reduce video data rate on the premise of visual effect as much as possible. During the compression of image, to get a high compressed ratio,people always only get a degraded image. This article aims to estimate the degraded model of the video compresed image through using the neural network and H.264 rate-distortion model.we can improve the quality of the compressed image, using neural network's predictive ability strongly and high fault-tolerance, improving the quantization parameter prediction in the video encoding, completing video image reconstruction. The experiments show that the neural network can improve the image quality of video compressed image greatly.