Research On JPEG-Like Image Compression And Hevc Video Coding&Watermarking Algorithms Based On APBT

With the rapid development of multimedia technology and network communication,people can easily access high-definition picture and video information through the internet.The transmission of multimedia data is also increasing the pressure of the CPU and network,at the same time,the transmission of the network also brings the problem of copyright protection,and authentication of digital products such as image and video.To solve the above problems,firstly,the background,significance and the current situation of image and video compression standard,parallel computing based on Graphics Processing Unit(GPU),and video watermarking techniques are described.Then,the parallel image compression coding based on GPU and High Efficiency Video Coding(HEVC)is studied,the related technology involved in HEVC is deeply analyzed,and the improvement of its performance is explored.At the same time,combined with the new generation of video compression standard HEVC,the preprocessing scheme of the watermark and the embedding scheme of the video watermark are studied in depth.The main contents of this thesis are as follows:(1)The JPEG image coding standard is analyzed in depth.According to the performance bottleneck of JPEG image coding,parallel JPEG image coding based on GPU is proposed,which can improve the efficiency of nearly 100 times compared with the conventional JPEG image coding scheme.(2)To solve the blocking problem caused by the block-based Discrete Cosine Transform(DCT)in conventional JPEG coding at low bit rates,combined with Discrete Sine Transform(DST)and the All Phase Digital Filtering(APDF)theory,the All Phase Discrete Sine Biorthogonal Transform(APDSBT)is proposed,which further expands the All Phase Biorthogonal Transform(APBT)theory.And apply it to JPEG encoding to replace the conventional DCT transform,and realize the JPEG coding based on parallel APDSBT.The parallel APDSBT-JPEG image coding algorithm improves the efficiency of the algorithm while solving the block effect caused by the conventional JPEG coding at low bit rate,which improves the subjectivity and objective performance compared with the conventional JPEG coding.(3)This thesis deeply analyzes the video coding framework of HEVC,and proposes the HEVC video coding algorithm based on APBT transform,and uses APBT transform to replace the DCT transform in the original standard.By analyzing the effect of APBT transform on the HEVC video coding performance of different Coding Unit(CU),a set of optimal replacement schemes is proposed,and the integer APBT transform is deduced,the relationship between APBT and DCT is studied,and the coding performance of HEVC is improved without increasing the complexity of the algorithm.(4)A video watermarking algorithm based on HEVC compression domain is proposed.The scheme chooses to embed watermark information in the quantized coefficients.The algorithm can solve the cumulative error caused by watermark embedding,and APBT transform is used to compress the watermark data in the digital watermark preprocessing stage,optimize the watermark embedding scheme,and the capacity of video watermark is improved.Due to the characteristics of HEVC video coding standard,HEVC video watermarking based on compressed domain cannot resist re-compression attack.Therefore,this thesis proposes a video watermarking algorithm based on APBT and Singular Value Decomposition(SVD).The proposed algorithm can resist the HEVC compression and showing excellent stability.