Underwater image enhancement web app powered by LU2Net, running entirely in the browser via ONNX Runtime Web.
Live demo: https://alenaksu.github.io/aquatune/
The neural network used in this project is LU2Net (Lightweight Underwater UNet), developed by Haodong Yang, Jisheng Xu, Zhiliang Lin, and Jianping He at Shanghai Jiao Tong University.
Yang, H., Xu, J., Lin, Z., & He, J. (2024). LU2Net: A Lightweight Network for Real-time Underwater Image Enhancement. arXiv:2406.14973. https://arxiv.org/abs/2406.14973
Pre-trained weights and source code: https://github.com/MrYangHaodong/LU2Net
Training uses the LSUI (Large Scale Underwater Image) dataset:
Peng, L., Zhu, C., & Bian, L. (2023). U-Shape Transformer for Underwater Image Enhancement. IEEE Transactions on Image Processing.
Dataset download: https://www.kaggle.com/datasets/noureldin199/lsui-large-scale-underwater-image-dataset