The Improvement Algorithm Of HEVC Screen Content Coding Based On Text-Block Recognition And Non-Square Hash-Based Blocks Matching

High Efficiency Video Coding(HEVC)is the current video compression standard.In comparison to Advanced Video Coding(AVC),HEVC offers about double the data compression ratio at the same level of video quality.The Screen Content Coding(SCC),as an extremely important extension of HEVC,are applied in wireless display,remote computer desktop access,and real-time screen sharing.It is because that several coding schemes have been proposed for SCC.These schemes have exerted outstanding coding performance.Which include intra block copy(IBC),palette mode,adaptive color transform(ACT),adaptive motion vector resolution(AMVR),and hash-based block matching.SCC presents a more time complexity along with outstanding performance compred with HEVC.For text sequences,this paper presented fast coding unit(CU)partition algorithm for alleviating the computational complexity of the SCC intra encoder based on analysing the coding character of intra direction prediction and intra block copy.Coding blocks are classified by text block and non-text block,which also have different coding process by the classification.Experiment results shows the proposed algorithm enabled the reduction of encoding time by up to 23%and a BD-rate increase by 0.43%with all intra configuration.SCC uses a hash-based block matching scheme to improve compression efficiency of screen content,particularly for text content.However,in this scheme,the available prediction block sizes only include square blocks.It means that the coding performance can be further improved.Thus,in this study,we analyzed the complication of building a hash table and through building 4×8 and 8×4 hash-based block matching for inter-picture prediction,the proposed scheme achieves 0.6%and 0.3%BD-rate saving with low delay and random access configurations with negligible encoding time increase for all test sequences.