Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

QD-PCQA: Quality-Aware Domain Adaptation for Point Cloud Quality Assessment

Guohua Zhang, Jian Jin, Meiqin Liu, Chao Yao, Weisi Lin

TL;DR

A novel Quality-aware Domain adaptation framework for PCQA, termed QD-PCQA, which includes a Rank-weighted Conditional Alignment strategy that aligns features under consistent quality levels and adaptively emphasizes misranked samples to reinforce perceptual quality ranking awareness.

Abstract

No-Reference Point Cloud Quality Assessment (NR-PCQA) still struggles with generalization, primarily due to the scarcity of annotated point cloud datasets. Since the Human Visual System (HVS) drives perceptual quality assessment independently of media types, prior knowledge on quality learned from images can be repurposed for point clouds. This insight motivates adopting Unsupervised Domain Adaptation (UDA) to transfer quality-relevant priors from labeled images to unlabeled point clouds. However, existing UDA-based PCQA methods often overlook key characteristics of perceptual quality, such as sensitivity to quality ranking and quality-aware feature alignment, thereby limiting their effectiveness. To address these issues, we propose a novel Quality-aware Domain adaptation framework for PCQA, termed QD-PCQA. The framework comprises two main components: i) a Rank-weighted Conditional Alignment (RCA) strategy that aligns features under consistent quality levels and adaptively emphasizes misranked samples to reinforce perceptual quality ranking awareness; and ii) a Quality-guided Feature Augmentation (QFA) strategy, which includes quality-guided style mixup, multi-layer extension, and dual-domain augmentation modules to augment perceptual feature alignment. Extensive cross-domain experiments demonstrate that QD-PCQA significantly improves generalization in NR-PCQA tasks. The code is available at https://github.com/huhu-code/QD-PCQA.

QD-PCQA: Quality-Aware Domain Adaptation for Point Cloud Quality Assessment

TL;DR

A novel Quality-aware Domain adaptation framework for PCQA, termed QD-PCQA, which includes a Rank-weighted Conditional Alignment strategy that aligns features under consistent quality levels and adaptively emphasizes misranked samples to reinforce perceptual quality ranking awareness.

Abstract

No-Reference Point Cloud Quality Assessment (NR-PCQA) still struggles with generalization, primarily due to the scarcity of annotated point cloud datasets. Since the Human Visual System (HVS) drives perceptual quality assessment independently of media types, prior knowledge on quality learned from images can be repurposed for point clouds. This insight motivates adopting Unsupervised Domain Adaptation (UDA) to transfer quality-relevant priors from labeled images to unlabeled point clouds. However, existing UDA-based PCQA methods often overlook key characteristics of perceptual quality, such as sensitivity to quality ranking and quality-aware feature alignment, thereby limiting their effectiveness. To address these issues, we propose a novel Quality-aware Domain adaptation framework for PCQA, termed QD-PCQA. The framework comprises two main components: i) a Rank-weighted Conditional Alignment (RCA) strategy that aligns features under consistent quality levels and adaptively emphasizes misranked samples to reinforce perceptual quality ranking awareness; and ii) a Quality-guided Feature Augmentation (QFA) strategy, which includes quality-guided style mixup, multi-layer extension, and dual-domain augmentation modules to augment perceptual feature alignment. Extensive cross-domain experiments demonstrate that QD-PCQA significantly improves generalization in NR-PCQA tasks. The code is available at https://github.com/huhu-code/QD-PCQA.
Paper Structure (24 sections, 12 equations, 3 figures, 7 tables)

This paper contains 24 sections, 12 equations, 3 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Point Cloud Quality Assessment
  4. Unsupervised Domain Adaptation
  5. Method
  6. Data Preprocessing Module
  7. QFA Strategy
  8. QSM Module
  9. Multi-Layer Extension Module
  10. Dual-Domain Augmentation Module
  11. RCA Strategy
  12. Quality Prediction Module
  13. Training Strategy
  14. Experiments
  15. Experimental Setup
  16. ...and 9 more sections

Figures (3)

  • Figure 1: Comparison with IT-PCQA. IT-PCQA performs feature alignment via UDA, overlooking the characteristics of quality perception. In contrast, our QD-PCQA introduces two quality-aware strategies. The RCA strategy aligns features guided by quality scores of two domains, promoting quality-consistent adaptation. The QFA strategy enriches feature diversity through a multi-layer QSM and SM. Additionally, a rank-weighted module emphasizes misranked feature pairs to mitigate ranking bias.
  • Figure 2: Architecture of the proposed QD-PCQA. Given the source-domain image $x_s$ and target-domain point cloud $x_t$, both are cropped to the same size. Then processed by QFA for quality-guided feature augmentation and by RCA for rank-aware conditional alignment. Finally, the predictor outputs the final quality score.
  • Figure 3: t-SNE visualization of the aligned representations learned from the source and target domains. Label values are denoted by color gradients. Ten label values are selected from the range of the variable for visualization.