{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/attention-guided-network-for-ghost-free-high","title":"Attention-guided Network for Ghost-free High Dynamic Range Imaging","arxiv_id":"1904.10293","date":"2019-04-23","proceeding":"CVPR 2019 6","authors":["Qingsen Yan","Dong Gong","Qinfeng Shi","Anton Van Den Hengel","Chunhua Shen","Ian Reid","Yanning Zhang"],"abstract":"Ghosting artifacts caused by moving objects or misalignments is a key\nchallenge in high dynamic range (HDR) imaging for dynamic scenes. Previous\nmethods first register the input low dynamic range (LDR) images using optical\nflow before merging them, which are error-prone and cause ghosts in results. A\nvery recent work tries to bypass optical flows via a deep network with\nskip-connections, however, which still suffers from ghosting artifacts for\nsevere movement. To avoid the ghosting from the source, we propose a novel\nattention-guided end-to-end deep neural network (AHDRNet) to produce\nhigh-quality ghost-free HDR images. Unlike previous methods directly stacking\nthe LDR images or features for merging, we use attention modules to guide the\nmerging according to the reference image. The attention modules automatically\nsuppress undesired components caused by misalignments and saturation and\nenhance desirable fine details in the non-reference images. In addition to the\nattention model, we use dilated residual dense block (DRDB) to make full use of\nthe hierarchical features and increase the receptive field for hallucinating\nthe missing details. The proposed AHDRNet is a non-flow-based method, which can\nalso avoid the artifacts generated by optical-flow estimation error.\nExperiments on different datasets show that the proposed AHDRNet can achieve\nstate-of-the-art quantitative and qualitative results.","url_abs":"http://arxiv.org/abs/1904.10293v1","url_pdf":"http://arxiv.org/pdf/1904.10293v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"attention-guided-network-for-ghost-free-high","repo_url":"https://github.com/JimmyChame/The-State-of-the-Art-in-HDR-Deghosting","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"attention-guided-network-for-ghost-free-high","repo_url":"https://github.com/Pea-Shooter/ADNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"attention-guided-network-for-ghost-free-high","repo_url":"https://github.com/drhdr-user/drhdr","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"attention-guided-network-for-ghost-free-high","repo_url":"https://github.com/liuzhen03/ADNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"attention-guided-network-for-ghost-free-high","repo_url":"https://github.com/qingsenyangit/AHDRNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"optical-flow-estimation","task_name":"Optical Flow Estimation"},{"task_slug":"high","task_name":"Vocal Bursts Intensity Prediction"}],"methods":[{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-block","method_name":"Dense Block"},{"method_slug":"relu","method_name":"ReLU"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1904.10293","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}