Fast Inter Prediction Encoding Based on Occupancy Map in V-PCC

In Video-based Point Cloud Compression (V-PCC), the point cloud system converts the input 3D image into multiple 2D images to represent the original data. These 2D images include a geometry map, a texture map, and an occupancy map. The occupancy map consists only of binary values, 0 and 1, to indicate whether a pixel value exists at a specific location in the input 3D point cloud data. Currently, the High-Efficiency Video Coding (HEVC) codec, designed for natural image compression and decompression, is used in this process. However, the occupancy map's characteristics cannot be fully leveraged within HEVC. In this paper, we propose a fast inter-prediction method that takes advantage of the binary nature of the occupancy map, consisting solely of 0s and 1s. During inter-prediction in HEVC, the codec evaluates multiple Coding Unit (CU) sizes to find the one with the lowest Rate-Distortion cost(RDcost). Our proposed algorithm enables the encoder to utilize the current occupancy map values to determine the best encoding mode for the CU block dynamically. This enables early termination of inter-prediction, leading to a significant reduction in overall encoding time. Our results demonstrate the effectiveness of this approach, achieving an encoding time reduction of up to 11.43%, thereby improving the overall efficiency of the encoding process.