Video Compression using Lower Coefficients Between Frames

Project Discription

Abstract - The broad use of videos and images now a days is popular. The advancement of communication in recent trends made video compression aimportantquality in transmission of information in social networking. Also the inadequate bandwidth for transmission and lower quality make video compression a serious phenomenon to consider in the field of communication. There is a need to improve the video compression process which can encode the video data with low computational complexity with better quality along with maintaining speed. The proposed video compression process is the one that is capable of removing all the redundancies in the video upto certain levels. The Discrete Cosine Transform Based block processing technique highly capable to keep the PSNR and gives better results.

Sparse Matrix

Here non zero position in the distance matrix need to be save in a matrix for future reference, when video need to be read. So sparse matrix is create for the frame which store non matching pixel value of the frames. Here one check is apply which help in identifying that either the new matrix size is more or less as compare to the previous one. If size is less then store in sparse matrix otherwise save original frame matrix. As it is already known that sparse matrix contain three column first represent row, second represent column and third represent value at that position.

Conclusions

As large number of video compression technique developed for different requirement. Proposed method is lossless video compression technique. Here work utilizes the frame base processing and un-similar information is store in sparse matrix of that frame. It has been obtain that proposed work is better as compare to the previous existing work on comparison of different parameters. It is also obtain that frame size affect the parameters like CR, PSNR. If frame size is small better compression ratio is obtained. As the frame size increase compression ratio is decrease.

Project Sample Image