RBFIM: Perceptual Quality Assessment for Compressed Point Clouds Using Radial Basis Function Interpolation

2025-03-18

AI Summary

Perceptual distortion assessment in point cloud compression (PCC) faces two key challenges: (1) conventional point-to-point nearest-neighbor matching fails to establish accurate, perceptually consistent correspondences; and (2) discrete features cannot effectively model geometry- and structure-related distortions that are salient to human vision. To address these, we propose a continuous feature modeling method based on radial basis function (RBF) interpolation, which maps discrete features of original and distorted point clouds onto smooth, continuous functions. This enables a unidirectional, bijective, and perceptually aligned coordinate-feature mapping—eliminating errors inherent in bidirectional search. To our knowledge, this is the first work to introduce RBF interpolation into point cloud quality assessment, yielding a subjectively driven continuous distortion metric. Evaluated on multiple subjective quality datasets, the proposed metric achieves over 12% improvement in Spearman rank-order correlation coefficient (SROCC) over PSNR and PCQM, significantly enhancing the perceptual feedback accuracy for codec optimization.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Feature Construction/Reformulation

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

One of the main challenges in point cloud compression (PCC) is how to evaluate the perceived distortion so that the codec can be optimized for perceptual quality. Current standard practices in PCC highlight a primary issue: while single-feature metrics are widely used to assess compression distortion, the classic method of searching point-to-point nearest neighbors frequently fails to adequately build precise correspondences between point clouds, resulting in an ineffective capture of human perceptual features. To overcome the related limitations, we propose a novel assessment method called RBFIM, utilizing radial basis function (RBF) interpolation to convert discrete point features into a continuous feature function for the distorted point cloud. By substituting the geometry coordinates of the original point cloud into the feature function, we obtain the bijective sets of point features. This enables an establishment of precise corresponding features between distorted and original point clouds and significantly improves the accuracy of quality assessments. Moreover, this method avoids the complexity caused by bidirectional searches. Extensive experiments on multiple subjective quality datasets of compressed point clouds demonstrate that our RBFIM excels in addressing human perception tasks, thereby providing robust support for PCC optimization efforts.

Problem

• Evaluate perceived distortion in point cloud compression.
• Improve accuracy of quality assessments using RBF interpolation.
• Address limitations of single-feature metrics in capturing human perception.

Innovation

• RBFIM uses radial basis function interpolation
• Converts discrete point features to continuous functions
• Improves accuracy in perceptual quality assessment