Full-reference Point Cloud Quality Assessment Using Spectral Graph Wavelets

Ryosuke Watanabe,Keisuke Nonaka,Eduardo Pavez,Tatsuya Kobayashi,Antonio Ortega

2024-06-14

Abstract:Point clouds in 3D applications frequently experience quality degradation during processing, e.g., scanning and compression. Reliable point cloud quality assessment (PCQA) is important for developing compression algorithms with good bitrate-quality trade-offs and techniques for quality improvement (e.g., denoising). This paper introduces a full-reference (FR) PCQA method utilizing spectral graph wavelets (SGWs). First, we propose novel SGW-based PCQA metrics that compare SGW coefficients of coordinate and color signals between reference and distorted point clouds. Second, we achieve accurate PCQA by integrating several conventional FR metrics and our SGW-based metrics using support vector regression. To our knowledge, this is the first study to introduce SGWs for PCQA. Experimental results demonstrate the proposed PCQA metric is more accurately correlated with subjective quality scores compared to conventional PCQA metrics.

Computer Vision and Pattern Recognition,Multimedia,Signal Processing

What problem does this paper attempt to address?

### What problems does this paper attempt to solve?

This paper aims to solve the challenges in **point - cloud quality assessment (PCQA)**, especially the problem of point - cloud quality degradation during processing, compression, and transmission in 3D applications. Specifically:

1. **Reliable quality assessment methods**: Reliable PCQA methods are crucial when developing compression algorithms with a good bit - rate - quality trade - off and techniques for improving quality (such as denoising).

2. **Limitations of existing methods**:
- Existing full - reference (FR) PCQA methods mainly focus on point - to - point errors and fail to fully consider more complex features (such as structural similarity or graph similarity).
- Learning - based FR - PCQA methods can be better correlated with subjective evaluation scores, but they have a long processing time and are less adaptable to different types of noise (such as Gaussian noise, down - sampling).

3. **Importance of multi - scale features**: Previous methods have not fully considered multi - scale features, which are very important for evaluating various types of distortion.

### Proposed solutions

To overcome the above problems, the paper proposes a full - reference point - cloud quality assessment (FR - PCQA) method based on **spectral graph wavelets (SGW)**. Specific contributions include:

1. **Introduction of SGW - PCQA metric**: A new PCQA metric is proposed by comparing the SGW coefficients of the coordinate and color signals between the reference point cloud and the distorted point cloud. This is the first study to use SGWs for PCQA.

2. **Combination of traditional and SGW - PCQA metrics**: Support vector regression (SVR) is used to combine the SGW - PCQA metric with two traditional FR - PCQA metrics to form an integrated version SGW - PCQA + to improve the correlation with subjective evaluation scores.

3. **Faster calculation speed**: Compared with traditional multi - scale FR - PCQA methods, SGW - PCQA uses fast calculation techniques (such as polynomial approximation) to significantly improve the calculation speed.

### Experimental results

The experimental results show that the proposed SGW - PCQA metric has a higher correlation with subjective quality scores, especially in the case of compression distortion and other types of noise (such as Gaussian noise, down - sampling). In addition, SGW - PCQA + outperforms existing PCQA methods on multiple datasets.

In conclusion, this paper improves the PCQA method by introducing SGWs, solves the shortcomings of existing methods in complex distortion evaluation, and provides a faster and more accurate quality assessment tool.