[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85243-en":3,"doc-seo-85243-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85243,1374391974564,"Clementine","https://ap-avatar.wpscdn.com/avatar/14000253aa45c000a9e?x-image-process=image/resize,m_fixed,w_180,h_180&k=1779874745381141002",8,"Research & Report","Water Reflection Detection Using Symmetric Attention","Water reflections challenge computer vision because standard deep models confuse real objects with their mirror images, leading to spurious false positives and negatives in detection and semantic segmentation. Detecting the reflection axis is essential for reliable scene understanding and robust object parsing. SAWRD-Net addresses this by leveraging imperfect reflective symmetry through dihedral group–equivariant convolutions and an end-to-end matrix-decomposition decoder, regressing axis keypoints and fitting the final axis via PCA.","arXiv :2607 . 10749v1 [ cs .CV] 12 Jul 2026  \nHighlights  \nWater Reflection Detection Using Symmetric Attention  \nShuxuan Yao, Chengjia Wang, Jianyuan Sun, Junyu Dong, Xinghui Dong  \n• We introduce an end-to-end water reflection detection network, i.e., SAWRDNet, which couples group-equivariant convolutions with low-rank matrix decomposition  \n• We develop a symmetric-attention mechanism that redistributes feature weights according to reflection cues, thereby improving the capture of symmetry-specific information and enhancing detection accuracy  \n• We design an MSRE block and an MD Decoder, which leverages the intrinsic symmetry of water reflection scenes while modeling long-range context via precise low-rank factorization  \nWater Reflection Detection Using Symmetric Attention  \nShuxuan Yaoa , Chengjia Wangb , Jianyuan Sunc , Junyu Donga , Xinghui Donga,∗  \na State Key Laboratory of Physical Oceanography and the Faculty of Information Science and Engineering, Ocean University of China, 238 Songling Road, Qingdao, 266100, Shandong, China  \nb School of Mathematical and Computer Sciences, Heriot-Watt University, Riccarton Mains Road, EH14  \n4AS, Edinburgh, United Kingdom  \nc School of Computer Science and Informatics, De Montfort University, Leicester, LE1 9BH, United  \nKingdom  \nAbstract  \nReflections of water pose a significant challenge for computer vision systems, as standard deep learning models frequently confuse objects with their mirror images, producing spurious false positives and negatives in tasks such as object detection and semantic segmentation. As a result, detecting reflection axes in natural-water scenes is pivotal for reliable object detection and scene understanding. To mitigate this issue, we leverage the intrinsic imperfect reflective symmetry of water and introduce a SymmetryAware Water Reflection Detection Network, namely, SAWRD-Net, that couples dihedral group–equivariant convolutions with a matrix-decomposition decoder in an end-toend framework. First, dihedral group convolutional layers extract geometry-consistent feature maps that explicitly encode both rotational and mirror symmetries. A Multiscale Reflection Equivariant block then aggregates features across scales and employs asymmetric-attention mechanism to highlight reflection-relevant regions. The proposed matrix-decomposition decoder factorizes high-dimensional features into compact lowrank parameter and confidence spaces, after which the network directly regresses keypoints on the reflection axis. Then a robust principal component analysis fits the final axis. Evaluated on the largest available water reflection scene data set, SAWRD-Net achieves a true-positive rate of 0.890 against human annotations, outperforming all  \n⋆ Code and models will be made available at [https:](https://github.com/INDTLab/SAWRD-Net)[//](https://github.com/INDTLab/SAWRD-Net)[github.com](https://github.com/INDTLab/SAWRD-Net)[/](https://github.com/INDTLab/SAWRD-Net)[INDTLab](https://github.com/INDTLab/SAWRD-Net)[/](https://github.com/INDTLab/SAWRD-Net)[SAWRD-Net](https://github.com/INDTLab/SAWRD-Net).  \n∗ Corresponding author  \nEmail addresses: [yaoshuxuan@stu.ouc.edu.cn](yaoshuxuan@stu.ouc.edu.cn) (Shuxuan Yao), [chengjia.wang@hw.ac.uk](chengjia.wang@hw.ac.uk)  \n(Chengjia Wang), [jianyuan.sun@dmu.ac.uk](jianyuan.sun@dmu.ac.uk) (Jianyuan Sun), [dongjunyu@ouc.edu.cn](dongjunyu@ouc.edu.cn) (Junyu Dong),  \n[xinghui.dong@ouc.edu.cn](xinghui.dong@ouc.edu.cn) (Xinghui Dong)  \nexisting water reflection detectors.  \nKeywords:  \nWater Reflection Detection, Symmetry Detection, Line Detection, Equivariant Learning, Matrix Decomposition.  \n1. Introduction  \nWater reflections are ubiquitous in nature photography and constitute a prominent class of outdoor images [1, 2, 3, 4], thus accurately modeling them benefits large-scale visual understanding [5] . Reflections of water create mirror-like duplicates of scene objects [6, 7] . Ideally, the object and its reflection would exhibit p","cbCaisfpYS1ZphwE","https://ap.wps.com/l/cbCaisfpYS1ZphwE","pdf",3163314,1,40,"English","en",105,"# Introduction\n## Problem of water reflection ambiguity\n## Limitations of feature-based methods\n## Prior deep learning approaches","[{\"question\":\"Why is water reflection detection difficult for computer vision models?\",\"answer\":\"Reflections create mirror-like duplicates that standard deep networks often treat as separate objects, causing false positives and negatives in detection and segmentation tasks.\"},{\"question\":\"What is SAWRD-Net designed to detect or estimate?\",\"answer\":\"SAWRD-Net targets the reflection axis in natural water scenes, separating real objects from their virtual mirrored counterparts.\"},{\"question\":\"How does SAWRD-Net improve symmetry capture and final axis estimation?\",\"answer\":\"It combines dihedral group–equivariant convolutions with a symmetric-attention mechanism, then uses a matrix-decomposition decoder to regress axis keypoints and applies robust PCA to fit the final axis.\"}]",1784202021,101,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"water-reflection-detection-using-symmetric-attention","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/water-reflection-detection-using-symmetric-attention/85243/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why is water reflection detection difficult for computer vision models?","Question",{"text":75,"@type":76},"Reflections create mirror-like duplicates that standard deep networks often treat as separate objects, causing false positives and negatives in detection and segmentation tasks.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is SAWRD-Net designed to detect or estimate?",{"text":80,"@type":76},"SAWRD-Net targets the reflection axis in natural water scenes, separating real objects from their virtual mirrored counterparts.",{"name":82,"@type":73,"acceptedAnswer":83},"How does SAWRD-Net improve symmetry capture and final axis estimation?",{"text":84,"@type":76},"It combines dihedral group–equivariant convolutions with a symmetric-attention mechanism, then uses a matrix-decomposition decoder to regress axis keypoints and applies robust PCA to fit the final axis.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":21,"slug":118},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]