Self-Refining Deep Symmetry Enhanced Network for Rain Removal
arXiv:1811.04761
Abstract
Rain removal aims to remove the rain streaks on rain images. The state-of-the-art methods are mostly based on Convolutional Neural Network~(CNN). However, as CNN is not equivariant to object rotation, these methods are unsuitable for dealing with the tilted rain streaks. To tackle this problem, we propose Deep Symmetry Enhanced Network~(DSEN) that is able to explicitly extract the rotation equivariant features from rain images. In addition, we design a self-refining mechanism to remove the accumulated rain streaks in a coarse-to-fine manner. This mechanism reuses DSEN with a novel information link which passes the gradient flow to the higher stages. Extensive experiments on both synthetic and real-world rain images show that our self-refining DSEN yields the top performance.
arXiv commentAccepted by ICIP 19. Corresponding author: Hanrong Ye
References in corpus (12)
- Adam: A Method for Stochastic Optimization
- Rotation-invariant convolutional neural networks for galaxy morphology prediction
- Pose Guided Person Image Generation
- Group Equivariant Convolutional Networks
- Rotation equivariant vector field networks
- Identity Mappings in Deep Residual Networks
- Exploiting Cyclic Symmetry in Convolutional Neural Networks
- Spherical CNNs
- Rotation Equivariance and Invariance in Convolutional Neural Networks
- Warped Convolutions: Efficient Invariance to Spatial Transformations
- Robustness of Rotation-Equivariant Networks to Adversarial Perturbations
- Deeply Supervised Rotation Equivariant Network for Lesion Segmentation in Dermoscopy Images