(A) study on overlapped block motion estimation and optimization-based video coding methods

Abstract

The main technical issue of low bit rate video communication is clearly the video coding method, which is required to achieve the necessary bit rate with sufficient image quality and reasonable hardware cost. Whereas most video coding standards uniquely define the bitstream syntax and the operation of the decoder, the exact nature of the encoder is generally left open to user specification. This implies that, given decoder configuration and bit rate, the quality of the encoded video depends on the strategy employed in the video encoder. Therefore, how to optimize the encoding algorithms effectively is the key factor deciding the value of the final product of the video codec system. In this dissertation, we propose effective video encoding methods improving the overall quality of coded video in low bit rate coding. These include a motion estimation/compensation method and rate-distortion optimized video coding methods. Overlapped block motion compensation(OBMC) has been proposed as a partial solution to the block artifact problems of conventional non-overlapped block motion compensation(BMC). OBMC has been shown to provide improved prediction efficiency as well as better subjective quality compared with BMC. Also, by using motion estimation techniques optimized for OBMC, called overlapped block motion estimation(OBME), we can obtain further substantial gains in compensation accuracy over conventional non-overlapped block motion estimation(BME). In this dissertation, first, theoretical analysis on OBMC is presented. Previous studies have demonstrated that the performance of OBMC is better than that of BMC. However, most of them are extensively based on the exhaustive experiments using some test sequences. Due to this lack of theoretical study, the properties of OBMC have not yet been fully understood. In this work, we theoretically evaluate the performance of OBMC compared with BMC, based on the one-dimensional signal model characterized by the AR(1) process an...