Multi-stream POINTNet-based Model For Blind Geometric Point Cloud Quality Assessment
Résumé
The evaluation of 3D point cloud quality is a critical com- ponent in the development of immersive multimedia systems for real-world applications. While perceptual quality evalua- tion techniques for 2D images and videos have reached high performances, developing robust and efficient blind metrics for point cloud quality assessment is still challenging. In this paper, we propose a no-reference point cloud quality as- sessment method that evaluates the quality of degraded 3D objects using an end-to-end multi-stream point-based model. We introduce geometric coordinates, normals, and curvatures as inputs to the proposed model stream to extract significant degradation features that are incorporated to predict the vi- sual quality of 3D point clouds. Experimental results on two benchmark databases: ICIP202 and SJTU, demonstrate that the proposed model achieves promising performance com- pared to state-of-the-art full and reduced methods.