Motion blur adaptive Rate Control

Abstract

In this paper, a new Rate Control (RC) algorithm is proposed to improve perceptual quality using motion blur estimation. The blurriness of captured video frames is estimated from a global motion vector, encoding frame rate and exposure time. For given blurriness of frames, the proposed RC algorithm reallocates coding bits between blurry frames and sharp frames. From the experimental results, the proposed RC algorithm achieves much higher quality after a motion blur region and similar quality in a motion blur region comparing to the RC algorithm in the JM reference software model.