Performance analysis of block matching criterion in video data on embedded processor using VHDL

Video Compression has played an important role in Multimedia data storage and transmission. Video compression techniques remove spatial as well as temporal redundancy using intra-frame and inter-frame coding respectively. A large level of compression can be achieved through inter-frame coding. In this paper, performance of four matching criterion in the temporal coding of video signal, which are Minimum Mean Absolute Error, Vector Matching Criterion, Smooth Constrained - Mean Absolute Error, and proposed algorithm using new pixel values are compared on hardware design using VHDL. Three step search algorithm has been used for searching the matching block as a block matching technique, because three step search algorithm is very simple and efficient search algorithm, which provides near optimum results only in three steps. For the various videos it has been experimentally observed that the minimum average error per pixel, and minimum search points per block reduces up to 0 per pixel and 13.31 per block respectively using proposed criterion and the values for the same parameters and same set of frames are very high using other criterion.