CNN-Based Cross-Dataset No-Reference Image Quality Assessment

Dan Yang, Veli-Tapani Peltoketo, Joni-Kristian Kämäräinen

ICCVW 2019 pp. 3913-3921

doi:10.1109/ICCVW.2019.00485 /iccvw/2019/yang2019iccvw-cnnbased/

Abstract

Recent works on no-reference image quality assessment (NR-IQA) have reported good performance for various datasets. However, they suffer from significant performance drops in cross-dataset evaluations which indicates poor generalization power. We propose a Siamese architecture and training procedures for cross-dataset deep NR-IQA that achieves clearly better performance. Moreover, we show that the architecture can be further boosted by i) pre-training with a large aesthetics dataset and ii) adding low-level quality cues, sharpness, tone and colourfulness, as additional features.

ICCVW Semantic Scholar

Cite

Text

Yang et al. "CNN-Based Cross-Dataset No-Reference Image Quality Assessment." IEEE/CVF International Conference on Computer Vision Workshops, 2019. doi:10.1109/ICCVW.2019.00485

Markdown

[Yang et al. "CNN-Based Cross-Dataset No-Reference Image Quality Assessment." IEEE/CVF International Conference on Computer Vision Workshops, 2019.](https://mlanthology.org/iccvw/2019/yang2019iccvw-cnnbased/) doi:10.1109/ICCVW.2019.00485

BibTeX

@inproceedings{yang2019iccvw-cnnbased,
  title     = {{CNN-Based Cross-Dataset No-Reference Image Quality Assessment}},
  author    = {Yang, Dan and Peltoketo, Veli-Tapani and Kämäräinen, Joni-Kristian},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2019},
  pages     = {3913-3921},
  doi       = {10.1109/ICCVW.2019.00485},
  url       = {https://mlanthology.org/iccvw/2019/yang2019iccvw-cnnbased/}
}