Point Cloud Quality Assessment Using the Perceptual Clustering Weighted Graph (PCW-Graph) and Attention Fusion Network

TL;DR

This paper introduces a novel no-reference point cloud quality assessment method that leverages perceptual clustering, weighted graph modeling, and attention fusion networks to evaluate 3D content quality without reference models.

Contribution

It proposes a new NR-PCQA approach combining perceptual clustering, weighted graph features, and attention-based fusion for improved accuracy.

Findings

Achieves higher correlation with human judgments compared to existing methods.

Demonstrates robustness across diverse point cloud datasets.

Outperforms state-of-the-art NR-PCQA techniques.

Abstract

No-Reference Point Cloud Quality Assessment (NR-PCQA) is critical for evaluating 3D content in real-world applications where reference models are unavailable.