zzc-1998 / nr-3dqa

Point cloud version for "No-Reference Quality Assessment for 3D Colored Point Cloud and Mesh models".

Paper

This is the code for "No-Reference Quality Assessment for 3D Colored Point Cloud and Mesh Models" and it is the point cloud version. This paper has been accepted by IEEE Transactions on Circuits and Systems for Video Technology. The arxiv version can be found here [http://arxiv.org/abs/2107.02041] and the TCVST version can be found here [https://ieeexplore.ieee.org/document/9810024].

How to start with the code?

You should get the h5py, pyntcloud, skimage package by

pip install h5py
pip install pyntcloud
pip install scikit-image

Environment settings

We test the code with Python 3.7 (and higher) on the Windows platform and the code may run on linux as well.

Demo

The demo.py includes a demo for the feature extraction of colored point cloud and we provide with a colored point cloud sample hhi_5.ply.

Database

The code is tested on the SJTU-PCQA database which can be downloaded at [https://smt.sjtu.edu.cn/].

The code is also tested on the WPC database which can be downloaded at [https://github.com/qdushl/Waterloo-Point-Cloud-Database].

Experiment Update

We update the experiment files for SJTU-PCQA and WPC databases, which includes the MOSs and extracted features from point clouds. We do not use the GGD, AGGD, Gamma parameters of color features in this experiment version for simplification.

Other PCQA works

We implement and collect several common PCQA metrics, which can be accessed here.

Citation

If you find our work useful, please cite our work as:

@ARTICLE{zhang2022no,
  author={Zhang, Zicheng and Sun, Wei and Min, Xiongkuo and Wang, Tao and Lu, Wei and Zhai, Guangtao},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={No-Reference Quality Assessment for 3D Colored Point Cloud and Mesh Models}, 
  year={2022},
  volume={},
  number={},
  pages={1-1},
  doi={10.1109/TCSVT.2022.3186894}}

If you have further questions, please email us through [email protected].

nr-3dqa's Issues

Feature Vector Dimensions Mismatch

Hello,

I was evaluating your code. The paper says it gives 88 feature values for 3D Point Clouds but feature_extract.py returns 64 values feature vector. Can you please explain the reason why is it doing so?
Thank you

Computing features.csv and mos.csv for custom dataset

Hello,

I was trying to test the method on my custom point cloud to compute SRCC, PLCC and KRCC. How can I compute features.csv and mos.csv. I guess features.csv can be obtained from demo.py with little modifications(please confirm) but mos.csv is the major issue. Please share some more information for custom point clouds. Thank you

Enquiry About Mesh Version of No-Reference Evaluation Method

Hi, I recently came across your paper on the no-reference evaluation method for point cloud models. Thanks for your great work!

I am currently working on a project that involves evaluating both point cloud and mesh models, and I believe your approach could be highly beneficial for my project. Yet I noticed that this repository only includes the point cloud version. Would it be possible to share a similar repository for the mesh version of your no-reference evaluation method?

Thank you for your time and consideration.
Best,