Improving Submesh Boundary Errors in Dynamic Mesh Coding

The Video-based Dynamic Mesh Coding (V-DMC) standard is a cutting-edge technology for compressing dynamic mesh data. V-DMC realizes parallel and partial decoding by introducing submesh frameworks, where dynamic meshes are segmented and processed independently. However, V-DMC may introduce submesh boundary errors due to misalignments in the existence or coordinates of vertices. Here, submesh boundary errors are defined as the misalignments of vertex coordinates and connectivity between adjacent submesh boundaries. These errors can create holes, thereby degrading both the objective and subjective quality of the decoded dynamic meshes. To minimize boundary errors and enhance coding performance, we propose a two-stage boundary error correction method integrated into V-DMC's preprocessing and encoding/decoding processes. Specifically, the first stage involves rearranging the preprocessing order to minimize boundary errors, and the second stage fills holes using boundary information. Experimental results demonstrate that the proposed method effectively minimizes boundary errors in V-DMC decoded meshes, significantly improving both objective and subjective quality compared to the V-DMC reference software. The D1 BD-Rate gain, indicative of coding efficiency, averages -16.3% for all intra and -25.9% for random access conditions.