Seamlessly run quality assessment experiments on subjective datasets
QualityLIB is a library that simplifies running quality assessment experiments on video datasets in Python. QualityLIB library interfaces with the VideoLIB package to provide an easy API that simplifies quality assessment research tasks such as
- Specifying and reading datasets of videos, conforming to various ITU standards.
- Standardizing the implementation of quality models using the
FeatureExtractor
class. - Simplifying the execution of feature extraction over datasets using the
Runner
class. - Standardizing the results of quality modeling using the
Result
class. - Easy interfacing with Scikit-Learn regressor models for routines such as
cross_validation
.
Refer to the official documentation for examples using QualityLIB and detailed descriptions of the API.
To use QualityLIB, you will need Python >= 3.7.0. To install using pip
, run
pip install git+https://github.com/abhinaukumar/qualitylib@main
To install using conda
, install pip
and git
in your environment using
conda install git pip
and use the pip
command above.
The goal of QualityLIB is to share with the community a tool that I build to accelerate my own quality assessment research workflows, and one that I have found great success with. Any feedback that can improve the quality of QualityLIB for the community and myself is greatly appreciated!
Please file an issue if you would like to suggest a feature, or flag any bugs/issues, and I will respond to them as promptly as I can. Contributions that add features and/or resolve any issues are also welcome! Please create a pull request with your contribution and I will review it at the earliest.
If you would like to contact me personally regarding QualityLIB, please email me at either abhinaukumar@utexas.edu or ab.kumr98@gmail.com.
QualityLIB is covered under the MIT License, as shown in the LICENSE file.