Application of TV image compression technology based on neural network

Tao Guo, Zhengqi Liu 

School of information engineering, Longdong University, 745000, China

Aiming at the disadvantages of digital TV, including a large amount of information and redundant information, a method of TV image compression technology based on neural network combining neural network with image compression technology is proposed in the work. Firstly, the TV image is divided into blocks as the input of neural network to build the network; secondly, the blocks are rebuilt to realize image compression recovery. The simulations show that the neural network algorithm can achieve the TV image compression effectively and the number of neurons of the hidden layer based on the neural network algorithm has great influence on the building and training of the network by contrast. When the number of neurons of the hidden layer is less, the image compression ratio will be higher and the image compression quality will be lower.