Benchmarking of DNN-based and L/VLM-based no-reference video quality models for full-HD and 4K video quality assessment (Bachelor)

In the last years, there has been a paradigm shift in the development of video quality model with machine learning and deep learning approaches gaining prominence. The models using these approaches are shown to outperform the state-of-the-art models developed using traditional signal-based approaches by the respective developers. However, the performance of these of larger pool of datasets is rarely investigated. In this thesis, multiple DNN-based and L/VLM-based video quality models with focus on no-reference models will be benchmarked for the case of full-HD and 4K videos quality assessment. The tasks would include

Collection and curation of relevant datasets
Choosing appropriate models
Video quality score generation using the selected models
Comparison of performance of different models and systematic analysis of the differences in comparison

Relevant literature (non-exhaustive)

M. H. Pinson, "Why No Reference Metrics for Image and Video Quality Lack Accuracy and Reproducibility," in IEEE Transactions on Broadcasting, vol. 69, no. 1, pp. 97-117, March 2023, doi: 10.1109/TBC.2022.3191059
Wu, H., Zhang, E., Liao, L., Chen, C., Hou, J., Wang, A., ... & Lin, W. (2023). Exploring video quality assessment on user generated contents from aesthetic and technical perspectives. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 20144-20154), https://doi.org/10.48550/arXiv.2211.04894
C. He, Q. Zheng, R. Zhu, X. Zeng, Y. Fan and Z. Tu, "COVER: A Comprehensive Video Quality Evaluator," 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA, 2024, pp. 5799-5809, doi: 10.1109/CVPRW63382.2024.00589.
Wu, H., Zhang, Z., Zhang, W., Chen, C., Liao, L., Li, C., ... & Lin, W. (2023). Q-align: Teaching lmms for visual scoring via discrete text-defined levels. arXiv preprint arXiv:2312.17090., https://doi.org/10.48550/arXiv.2312.17090
https://github.com/google/uvq

Suitable for Bachelor theses. Contact person: Dr.-Ing. Rakesh Rao Ramachandra Rao