Evaluation for random access performance of Multi-view View Coding

For the Multi-View Video Coding (MVC), improving coding efficiency is the primary target, but supporting random access is also an important requirement. Generally, the number of frames needed to be decoded can be used to evaluate the random access performance. However the most intuitive way to evaluate the random access performance is the random response time depending on the size of the actual number of transmitted bits. In addition, the output bits number of the coded frame can be derived from its prediction efficiency, so in this paper, making use of the prediction efficiency, we propose a method to evaluate the random access performance which use the output bits number as measurement. Experimental results show that the evaluation results are more accurate. For the prediction efficiency can also be used to estimate the compressing efficiency, this method plays an important role on looking for an optimal coding scheme which balances coding efficiency and random access performance reasonably.