Study Of Video Enhancement Algorithms For HDTV

HDTV (High Definition Television) reflects the future of TV for its high resolution and large field of vision, which has attracted much more attention of numerous merchants and researchers. With the broadcasting of HDTV signal, there exists the mismatch of display rate in different displays. For this purpose, this paper put the emphasis on the frame rate up-conversion algorithm.Firstly, a novel concept of motion targets is presented. Using this concept to replace the traditional block mode, a novel frame rate up-conversion algorithm based on motion targets is developed. It includes four steps: motion object segmentation, motion targets segmentation, motion estimation, motion compensation. For the first step, a morphology operator-based video motion object segmentation algorithm is proposed. By using it, one can obtain completed motion objects. Then, a multiscale morphology gradient-based watershed segmentation algorithm is introduced to segment the video frame in the spatial domain. Combined with motion objects, motion targets can be obtained. And a motion estimation algorithm is developed based on motion targets mode. The final step is motion compensation. Each frame is partitioned into several portions. For different portion, different motion compensation methods are adopted. Experimental results illustrate that the proposed frame rate up-conversion algorithm can obtain very good subjective vision quality.