MPEG-4 "Block DCT+Shape Adaptive DCT" Algorithm Research And Its FPGA Hardware Implementing

MPEG-4 is a new ISO/IEC Audio/visual coding standard developed by MPEG (Moving Picture Experts Group). It is different from its early MPEG-1/2 standards and is coded /decoded based on video objects of arbitrary shape. The content based interactive concept then becomes the key technology of MPEG-4, which is leading to a new development for visual coding and multimedia application in the future, and of course becomes an attracting research topic all over the world.After a detailed discuss of MPEG-4 image/video coding and decoding key technologies, this paper mainly researched into Shape-Adaptive DCT algorithm(which is widely used now in MPEG-4 media than DWT) and its properly FPGA implementing frame, including block DCT and SA-DCT modules.First we introduced the current video compression standards about their establishing background, main idea, base components, involve functions and their codec algorithms etc. and also discuss the video objects segmentation and forming, VOP (video objects plane) coding, scalable coding, moving estimate or compensation, sprite coding of MPEG-4. Then we aimed to the "block DCT + SA-DCT" transform algorithm to start the work, researched its FPGA hardware implementing, designed its modules with VHDL and accomplished simulation on Xilinx ISE 7.1i develop platform, and achieved the expected object with a correct waveform. Final we described the gains of this paper and planned the future research aims, which will helpful to understand the signification of "combining object based coding or decoding algorithms of MPEG-4 with FPGA technology" and begin more FPGA research work.