| Summary: | 碩士 === 國立成功大學 === 電機工程學系碩博士班 === 101 === This paper proposes a complexity control of motion compensation for the video encoder to choose suitable prediction mode that fits the hardware constraint of the decoder. In order to take the complexity into consideration, we combine the complexity term of motion compensation with the conventional rate-distortion optimization (RDO), and extend the equation to rate-distortion-complexity optimization (RDCO). Furthermore, we use a macroblock (MB) as a coding unit. In order to meet the target complexity, the Lagrange multiplier is updated for each coding unit to allocate appropriate computing complexity. In experimental result, the average control error is 1.16% under difference complexity and bitrate constrains for our proposed. It means our algorithm provides more accurate and stable control than reference. The average PSNR drop is 0.248 dB by reducing 50% complexity of motion compensation. Comparing to 1.167 dB PSNR drop of reference algorithm, our algorithm can provide better video quality.
|
|---|
