Semi-supervised Transfer Learning for Image Rain Removal

TL;DR

This paper introduces a semi-supervised transfer learning approach for single image rain removal, leveraging both synthetic and real rainy images to improve generalization and reduce bias in rain pattern learning.

Contribution

It proposes a novel semi-supervised learning paradigm that incorporates real rainy images without clean counterparts, enhancing model generalization to diverse rain types.

Findings

Outperforms state-of-the-art methods on synthetic data

Effectively generalizes to real rainy images

Reduces bias towards synthesized rain patterns

Abstract

Single image rain removal is a typical inverse problem in computer vision. The deep learning technique has been verified to be effective for this task and achieved state-of-the-art performance. However, previous deep learning methods need to pre-collect a large set of image pairs with/without synthesized rain for training, which tends to make the neural network be biased toward learning the specific patterns of the synthesized rain, while be less able to generalize to real test samples whose rain types differ from those in the training data. To this issue, this paper firstly proposes a semi-supervised learning paradigm toward this task. Different from traditional deep learning methods which only use supervised image pairs with/without synthesized rain, we further put real rainy images, without need of their clean ones, into the network training process. This is realized by elaborately formulating the residual between an input rainy image and its expected network output (clear image without rain) as a specific parametrized rain streaks distribution. The network is therefore trained to adapt real unsupervised diverse rain types through transferring from the supervised synthesized rain, and thus both the short-of-training-sample and bias-to-supervised-sample issues can be evidently alleviated. Experiments on synthetic and real data verify the superiority of our model compared to the state-of-the-arts.