Fast Intra-frame Coding Algorithm for HEVC Based on TCM and Machine Learning
MetadataShow full item record
High Efficiency Video Coding (HEVC) is the latest video coding standard. Compared with the previous standard H.264/AVC, it can reduce the bit-rate by around 50% while maintaining the same perceptual quality. This performance gain on compression is achieved mainly by supporting larger Coding Unit (CU) size and more prediction modes. However, since the encoder needs to traverse all possible choices to mine out the best way of encoding data, this large flexibility on block size and prediction modes has caused a tremendous increase in encoding time. In HEVC, intra-frame coding is an important basis, and it is widely used in all configurations. Therefore, fast algorithms are always required to alleviate the computational complexity of HEVC intra-frame coding. In this thesis, a fast intra-frame coding algorithm based on machine learning is proposed to predict CU decisions. Hence the computational complexity can be significantly reduced with negligible loss in the coding efficiency. Machine learning models like Bayes decision, Support Vector Machine (SVM) are used as decision makers while the Laplacian Transparent Composite Model (LPTCM) is selected as a feature extraction tool. In the main version of the proposed algorithm, a set of features named with Summation of Binarized Outlier Coefficients (SBOC) is extracted to train SVM models. An online training structure and a performance control method are introduced to enhance the robustness of decision makers. When applied on All Intra Main (AIM) full test and compared with HM 16.3, the main version of the proposed algorithm can achieve, on average, 48% time reduction with 0.78% BD-rate increase. Through adjusting parameter settings, the algorithm can change the trade-off between encoding time and coding efficiency, which can generate a performance curve to meet different requirements. By testing different methods on the same machine, the performance of proposed method has outperformed all CU decision based HEVC fast intra-frame algorithms in the benchmarks.