Content similarity based early skip mode decision for low power surveillance video compression

As a state-of-the-art video compression technique, H.264/AVC has been deployed in many surveillance cameras to improve the compression efficiency. However, it induces very high coding complexity, and thus high power consumption. In this paper, an early skip mode decision algorithm is proposed to reduce the computational complexity and power consumption in surveillance video compression by detecting the content similarity using chrominance and motion features. The overall computational complexity of the whole video encoding system is significantly reduced by 84%. No loss is observed in both of subjective and objective video quality.