Selective Compression of Video with H.264/AVC

X. Li, E.A. Edirisinghe, and H.E. Bez (UK)


H.264, shape coding, object based video coding, shape adaptive integer transform, video coding for surveillance, selective compression.


In this paper, we propose a novel scheme for H.264/AVC based selective compression of video for I-frames for applications in CCTV surveillance systems and bandwidth constraint scenarios. The proposal resolves a major limitation of H.264/AVC as compared to object based MPEG-4 video coding. The proposed scheme adopts a novel design of a Shape Adaptive Integer Transform to achieve the coding of the selected arbitrary shaped objects/content within a video frame. We propose the re-design of the slice group structure of H.264 video and means for encoding and decoding binary shape information of arbitrary shaped objects within H.264/AVC bit stream. The effects of the object coding on the texture representation and decoding methods also are discussed. Experimental results are provided to study the performance and flexibility of functionality of the proposed system. It’s possible application in CCTV surveillance systems are discussed.

Important Links:

Go Back