Zero-Shot Restoration of Back-lit Images Using Deep Internal Learning

Lin Zhang1, Lijun Zhang1, Xiao Liu1, Ying Shen1,*, Shaoming Zhang2, * and Shengjie Zhao1

1School of Software Engineering, Tongji University, Shanghai 201804, China
2College of Surveying and Geo-informatics, Tongji University, Shanghai 201804, China


This is the website of our paper "Zero-Shot Restoration of Back-lit Images Using Deep Internal Learning", published in ACM Multimedia 2019.

How to restore back-lit images still remains a challenging task. State-of-the-art methods in this field are based on supervised learning and thus they are usually restricted to specific training data. In this paper, we propose a "zero-shot" scheme for back-lit image restoration, which exploits the power of deep learning, but does not rely on any prior image examples or prior training. Specifically, we train a small image-specific CNN, namely ExCNet (short for Exposure Correction Network) at test time, to estimate the ``S-curve'' that best fits the test back-lit image. Once the S-curve is estimated, the test image can be then restored straightforwardly. ExCNet can adapt itself to different settings per image. This makes our approach widely applicable to different shooting scenes and kinds of back-lighting conditions. Statistical studies performed on 1512 real back-lit images demonstrate that our approach can outperform the competitors by a large margin. To the best of our knowledge, our scheme is the first unsupervised CNN-based back-lit image restoration method. To make the results reproducible, the source code is available on this website.

Source Code

1. ExCNet.ipynb

This is the code of ExCNet, including the network model and the back-lit images restoration procedure. The prerequisites for running the .ipynb file is the Tensorflow environment and Jupyter Notebook.


These are some back-lit images and videos for testing. The extract codes are 'v752' and 'yerr' respectively.

Created on: Jul. 10, 2019

Last update: Jul. 23, 2019