MPEG 포인트 클라우드 압축을 위한 차이값 고속 측정 방법

Abstract

In this thesis, we propose fast distortion quality measurement to improve efficient compression of massive point cloud data by introducing distortion search metric. In MPEG, point cloud compression standard is being actively standardized. As a new medium, point clouds pose new challenges in compression. One of them is to accurately assess the distortion of the encoded point clouds. If a point cloud is compressed lossy, there is no guarantee in mapping exact points between the original and encoded point clouds. For this reason, a tool called PC_ERROR [5] has been adopted for PCC. The PC_ERROR tool nicely solves the distortion measurement in an accurate manner; yet, it suffers from substantial computational complexity. To deal with this problem, we propose a fast distortion computing method to measure the Point Cloud Compression sequences. On top of PC_ERROR we designed new searching algorithm. Throughout the experimental results, we showed that the proposed method outperforms the PC_ERROR tool by 75 percent in execution time while preserving the accuracy in distortion measure.