CWCihang Wei

Cihang Wei

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Cihang Wei has a total of 10 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    MDeRainNet: An Efficient Macro-pixel Image Rain Removal Network

    by Tao Yan, Weilong Huang, Weijiang He, Chenglong Wang, Cihang Wei, Yiwei Lu, Xiangjie Zhu, Yinghui Wang, Rynson W. H. Lau

    Since raining weather always degrades image quality and poses significant challenges to most computer vision-based intelligent systems, image de-raining has been a hot research topic in computer vision community. Fortunately, in a rainy Light Field (LF) image, background obscured by rain streaks in one sub-view may be visible in the other sub-views, and implicit depth information and recorded 4D structural information may benefit rain streak detection and removal. However, existing LF image rain removal methods either do not fully exploit the global correlations of 4D LF data or only utilize partial sub-views (i.e., under-utilization of the rich angular information), resulting in sub-optimal rain removal performance and no-equally good quality for all de-rained sub-views. In this article, we propose an efficient neural network, called MDeRainNet , for rain streak removal from LF images. The proposed network adopts a multi-scale encoder–decoder architecture, which directly works on Macro-pixel Images (MPIs) for improving the rain removal performance. To fully model the global correlation between the spatial information and the angular information, we propose an Extended Spatial-angular Interaction (ESAI) module to merge the two types of information, in which a simple and effective Transformer-based Spatial-angular Interaction Attention (SAIA) block is also proposed for modeling long-range geometric correlations and making full use of the angular information. Furthermore, to improve the generalization performance of our network on real-world rainy scenes, we propose a novel semi-supervised learning framework for our MDeRainNet , which utilizes multi-level KL loss to bridge the domain gap between features of synthetic and that of real-world rain streaks and introduces colored-residue image-guided contrastive regularization to reconstruct rain-free images. Extensive experiments conducted on both synthetic and real-world Light Field Images (LFIs) demonstrate that our method outperforms the state-of-the-art methods both quantitatively and qualitatively.


  • SANet: An End-to-End Anaglyph Image Separation Network

    by Tao Yan, Lu Tan, Yiwei Lu, Cihang Wei, Jinsheng Liu, Rynson W. H. Lau

    Anaglyph image separation recovering a stereo image pair from an anaglyph image is a classical and ill-posed image recovery problem. We observed that there is always an interpenetration between the contents of the left and right views of an anaglyph image, as image compression and decompression are always adopted during the transmission of anaglyph images via Internet. Existing anaglyph image separation methods always produce stereo image pairs with obvious color degradation or structural distortion. In this paper, we propose an end-to-end network, called Stereo Image Separation from Anaglyph Images (SANet), consisting of two main parts: AIS-Net and DE-Net. The AIS-Net employs residual refinement module to separate the initial features extracted by VGG for reconstructing the left and right views. To address interpenetration between left and right views, we propose BiFIB, which separates residual information from one view and transfers it to another view. Since depth estimation from the recovered left and right views relies on the recovery of structural features for the left and right views, DE-Net is assumed to provide useful constraint to our SANet. Extensive experiments conducted on several representative datasets demonstrate the effectiveness and superiority of our method compared with existing methods. Our code and models are available at: https://github.com/YT3DVision/SANet