Integrated image color enhancement tool with image quality predictor
This project evaluates the task of refining the existing Deep Local Parametric Filters (DeepLPF) image enhancement tool, implemented by S. Moran et al. While the original model focuses on objective evaluation metrics for training, we have integrated a subjective evaluation metric into its training u...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
Nanyang Technological University
2024
|
| 主題: | |
| 在線閱讀: | https://hdl.handle.net/10356/174902 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 總結: | This project evaluates the task of refining the existing Deep Local Parametric Filters (DeepLPF) image enhancement tool, implemented by S. Moran et al. While the original model focuses on objective evaluation metrics for training, we have integrated a subjective evaluation metric into its training using an existing trained Neural Image Assessment (NIMA) tool, implemented by titu1994, as an image quality predictor. This project aims to find the combination of both models to produce the best results, by evaluating the integrated model based on the objective and subjective evaluation metrics to achieve a state of Pareto efficiency, where we find the optimal balance between both types of metrics. The results of the best integrated model are analysed, and the results are documented in this report. |
|---|