Fast Intra Mode Decision Algorithm of HEVC Based on Convolutional Neural Network

Aiming at the high complexity and large amount of computation of HEVC intra mode decision algorithm, a fast intra mode decision algorithm based on convolutional neural network is proposed. Firstly, different size of prediction units are scaled to the same size through bilinear interpolation, and then sent to convolution neural network to perform convolution pooling and other operations. Finally, the mode decision results are selected by softmax layer. Five standard test sequences with different resolutions are used to form data sets, and the convolutional neural network model is trained and tested. Experimental results show that the proposed algorithm is equivalent to the standard test software HM16.0 in image quality and video bit rate, but the total coding time is improved by about 20.14%.