HEVC (H.265) Intra-Frame prediction implementation Using MATLAB

Abstract: HEVC (H.265) standard is the latest enhanced video coding standard which was planned to improve the rendered specifications of its preceding standard MPEG-4 (H.264). According to the H.265 “The main goal of the HEVC standardization effort is to enable significantly improved compression performance relative to existing standards—in the range of 50% bit-rate reduction for equal perceptual video quality” [2]. Intra-picture prediction is a tool in HEVC which “uses some prediction of data spatially from region-to-region within a specific picture, but has no dependence on other pictures in the video frames” [2]. Intra-picture prediction of HEVC is the legacy of intra-frame prediction tool in H.264. Although both of them has the same approach in for spatial prediction of pictures based on spatial sample prediction followed by transform coding, H.265 intra-frame prediction uses much more developed features compared to H.264. An overview of the main features in intra-frame prediction of H.265 could be written as follows:  A quad-tree block division structure with respect to amount of details in an image  33 Angular modes in angular prediction (just 8 different modes in H.264)  Planar prediction for smoothing the sample surfaces [2] It’s worth mentioning that the quad-tree structure of H.265 intra normally uses square block with sizes in range 4, 8, 16, 32 and 64 (different block sizes based-on the level of granularity in the image), while in H.264 the processing units are up to macro-blocks of 16x16 samples. Moreover, while this video coding standard splits images to one luma and two chroma parts, thesis focuses only on the implementation of intra-prediction on luma part of an image. This thesis aims at implementation of the intra-frame prediction of HEVC using MATLAB. All the steps of implementation process are listed as follows:  Converting RGB images to YUV colour-space and working on the luma part (or Y)  Splitting images to square blocks ranging from 4 to 64 pixels  Implementing intra-frame prediction algorithm  Comparing intra-prediction output of H.264 and H.265 in square blocks with size 4 and 16 pixels This Thesis is organised in 3 main sections. The first and second sections revolve around literature review and definition of the concept of HEVC standard and intra-prediction respectively. The third section focuses on the implementation process and evaluation of the prediction algorithm. Finally, in the evaluation part, based-on statistical graphs derived from the output comparison of H.264 and H.265 intra-prediction for different images, it has been demonstrated that H.265 by far has a better image quality than of the H.264.