Low computation complexity and error-resilient video transmission for remote microscope monitoring system
Shengwei Pei, Minghui Du, Jianhua He, Quan Wan · 2005
This paper presents two video coding algorithms that are adopted in remote microscope monitoring system (RMMS). Due to translation motion of the specimen, all macro-blocks (MB) in a group of block (GOB) can share a common motion vector (CMV) that is the average of the first few MBs' motion vectors in the GOB. By using the CMV, the first algorithm reduces the times of searching matching MB significantly and improves the accuracy of motion estimation. In order to conceal the errors that are caused by specimen motion direction change and corrupted CMV, the second algorithm is proposed for error-control and data recovery based on the first algorithm. According to a criterion deduced from SAE (sum of absolute error), the second algorithm detects the MB in which the motion direction of the specimen is changed, and then the motion vector information can be recovered with the next GOB's and previous GOB's CMV by the second algorithm. The simulation results show that the first algorithm can reduce the computational load and improve the accuracy of motion estimation efficiently, and that the second algorithm has satisfactory performance in error-concealment.