Efficient Prediction Mode Decisions for Low Complexity MV-HEVC

The increased number of possible inter-prediction blocks sizes, improved the coding efficiency of High Efficiency Video Coding (HEVC), but at the same time it also increased the complexity of the encoder. In this work, we propose an Efficient Inter-Prediction Mode Decision (EIPMD) technique for the Multi-View extension of the HEVC standard (MV-HEVC). We figure out the least and most probable inter-prediction Modes and use Coding Unit (CU) splitting information of the base view to find its relation with the prediction Modes. We improve the efficiency of our technique by optimizing the base view search area. In our technique, we also makes use of the current depth level of the CU and temporal level of the pictures. Simulation results show that compared to HTM [16.2], which is the reference encoder for MV-HEVC, our proposed technique reduces the encoding time by approximately 66% on average, with very minor change in other parameters.