Towards a Point Cloud Structural Similarity Metric
read more
Citations
Predicting the Perceptual Quality of Point Cloud: A 3D-to-2D Projection-Based Exploration
Reduced Reference Perceptual Quality Model and Application to Rate Control for 3D Point Cloud Compression.
A Reduced Reference Metric for Visual Quality Evaluation of Point Cloud Contents
Mahalanobis Based Point to Distribution Metric for Point Cloud Geometry Quality Evaluation
Subjective Quality Database and Objective Study of Compressed Point Clouds with 6DoF Head-mounted Display
References
Image quality assessment: from error visibility to structural similarity
Geometric distortion metrics for point cloud compression
A Multiscale Metric for 3D Mesh Visual Quality Assessment
Watermarked 3-D Mesh Quality Assessment
Point Cloud Quality Assessment Metric Based on Angular Similarity
Related Papers (5)
Frequently Asked Questions (13)
Q2. What is the important metric in the MPCQA dataset?
The majority of features that capture uniformity of surface shape perform very poorly, with the exception of some metrics, namely, σ2, µAD and mAD, applied on curvature values.
Q3. What is the way to mitigate the quality scores of the features?
the authors propose the use of voxelization prior to feature extraction, in order to mitigate objective quality scores that are achieved at high-resolution models.
Q4. How many point clouds were used in this study?
The second dataset consists of 6 point clouds whose geometry was compressed using three codecs, namely, V-PCC, G-PCC (TriSoup module) and PCL, at three degradation levels.
Q5. What is the first dataset of point clouds?
The first dataset consists of 8 point clouds whose geometry and color were compressed1A prototype MATLAB implementation is made available online at https://www.epfl.ch/labs/mmspg/pointssim/.at 6 levels using five codecs, namely, V-PCC and the four GPCC test model variations.
Q6. What is the metric for evaluating the visual quality of 3D meshes?
Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Transactions on Image Processing, vol. 13, no.
Q7. What is the purpose of the analysis?
A Principal Component Analysis (PCA) is then issued to provide an orthonormal basis and a linear approximation of the local surface, which passes from the centroid of the neighborhood.
Q8. What are the common features extracted from a point cloud?
The features are extracted from computed quantities that depend on point cloud attributes, including geometry, normal vectors, curvature values, and colors.
Q9. What was the purpose of the study?
The point clouds were evaluated in three different sessions, with one of them being relevant to this study; that is a fixed-size point-based rendering with color information obtained from the original models after a re-coloring step.
Q10. What is the performance of the metrics for the MPCQA dataset?
It is evident that color-based features are over-performing, achieving high scores, with the best being a PCC of 0.928 and SROCC of 0.920, for σ2 and k = 12.
Q11. How is the performance of color-based features shown in the MPCQA dataset?
Based on their results, a remarkable performance increase is observed, with a maximum of 0.893 for PCC and 0.832 for SROCC at 8-bit voxel depth using mAD with k = 24.
Q12. What is the reason for the poor performance of the MPCQA dataset?
The general poor performance can be explained by the fact that (a) the dataset consists of several rather noisy point clouds, and (b) the original color values used for the decompressed models act as distractors.
Q13. What is the performance of color-based features in the MPCQA dataset?
In Figure 6a, the performance of color-based features is presented in the MPCQA dataset, after voxelization using bitdepths equal and below the lowest resolution among original models (i.e., this dataset consists of 9-bit and 10-bit models).