(b) depicts the key insight of our HDR deghosting approach with CA-ViT. As shown in Fig (a), the CA-ViT is designed as a dual-branch architecture where the global branch models long-range dependency among image contexts through a multi-head Transformer encoder, and the local branch explores both intra-frame local details and inner-frame feature relationship through a local context extractor. Extensive experiments on three benchmark datasets show that our approach outperforms state-of-the-art methods qualitatively and quantitatively with considerably reduced computational budgets. By incorporating the CA-ViT as basic components, we further build the HDR-Transformer, a hierarchical network to reconstruct high-quality ghost-free HDR images. For the local branch, we design a local context extractor (LCE) to capture short-range image features and use the channel attention mechanism to select informative local details across the extracted features to complement the global branch. Specifically, the global branch employs a window-based Transformer encoder to model long-range object movements and intensity variations to solve ghosting. The CA-ViT is designed as a dual-branch architecture, which can jointly capture both global and local dependencies. In this paper, we propose a novel Context-Aware Vision Transformer (CA-ViT) for ghost-free high dynamic range imaging. Restricted by the locality of the receptive field, existing CNN-based methods are typically prone to producing ghosting artifacts and intensity distortions in the presence of large motion and severe saturation. High dynamic range (HDR) deghosting algorithms aim to generate ghost-free HDR images with realistic details. 2022.07.04 Our paper has been accepted by ECCV 2022.2022.07.19 The source code is now available.2022.08.11 The arXiv version of our paper is now available.2022.08.26 The PyTorch implementation is now avaible.The PyTorch version is available at HDR-Transformer-PyTorch. This is the official MegEngine implementation of our ECCV2022 paper: Ghost-free High Dynamic Range Imaging with Context-aware Transformer ( HDR-Transformer). Ghost-free High Dynamic Range Imaging with Context-aware Transformerīy Zhen Liu 1, Yinglong Wang 2, Bing Zeng 3 and Shuaicheng Liu 3,1*ġMegvii Technology, 2Noah’s Ark Lab, Huawei Technologies, 3University of Electronic Science and Technology of China
0 kommentar(er)
