Firstly the original image is divided into several macro blocks in the light of certain size,and macro block within each row are permuted.

Motion vector of the moving object is obtained by one designed filter, which mainly processes motion vector of macroblockes of the moving object s boundary.

The function encode_one_macroblock() is the center in the implementation of the encoding algorithm.

This paper proposes a hierarchical model to describe the motion of each macroblock(MB)from coarsely to finely.

The improved APIB with the PSNR weighted to calculate macro-block complexity.

Base on the character that there is less temporal correlation between frames where scene change oc-curs,according to different macro-block prediction mode in the whole frame,the proposed algorithms obtain candidates of scene change frames.

This paper proposes a peer-to-peer framework, which adopts Skip graph named CredentialIndex as the overlay network for credential storage and credential chain discovery.

The proposed method can detect scene boundaries through using various macroblock type of B-frame and P-frame in MPEG video flow in order to realize accurate scene change positioning.

In this thesis,an adaptive macroblock encoding algorithm based on macroblock type spatial domain predic-tion is proposed.

The proposed method determines the motion activity and visual content change of the MPEG compressed video by using the luminance DCT DC value、macroblock type and motion vector.