Optimizing Projection-based Point Cloud Quality Assessment with Human Preferred Viewpoints Selection
How to select the best viewpoints for projection-based PCQA methods?
1Shanghai Jiaotong University, 2Nanyang Technological University
#Corresponding author.
The dataset can be downloaded via OneDrive.
The VP-PCQA.zip is formatted like:
|--dis_ply (All the distorted point clouds in .ply format.)
|--dis_vp_image |--AxeGuy_cn_20.ply |--0.png ~ 41.png
(The projection images of 42 viewpoints defined by the geodesic sphere for each distorted point cloud.)
|--dis_vp_json |--AxeGuy_cn_20.ply |--0.json ~ 41.json + vp_attribute.csv
(The camera parameters (.json) for the 42 viewpoints of each distorted point cloud, and preferred viewpoint index features in 'vp_attribute.csv'.)
|--ref_ply (All the reference point clouds in .ply format.)
|--ref_|--ref_ply
(The projection images of 42 viewpoints defined by the geodesic sphere for each reference point cloud.)
The human-selected viewpoints can be accessed here. (`human_preferred_index.csv' in this repo) The CSV file lists viewpoints in order of human preference, with the most favored viewpoints appearing first.
`cal_vp_attribute.py' can be used to calculate the preferred index features.
`vp_render.py' can be used to render the projections according to the geodesic sphere.
Please contact any of the first authors of this paper for queries.
- Zicheng Zhang,
zzc1998@sjtu.edu.cn, @zzc-1998
If you find our work interesting, please feel free to cite our paper:
@misc{zhang2024abench,
title={Optimizing Projection-based Point Cloud Quality Assessment with Human Preferred Viewpoints Selection},
author={Zicheng Zhang and Yu Fan and Wei Sun and Xiongkuo Min and Xiaohong Liu and Chunyi Li and Haoning Wu and Weisi Lin and Ning Liu and Guangtao Zhai},
year={2024},
booktitle={IEEE International Conference on Multimedia and Expo (ICME)}
}