{"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/srn-side-output-residual-network-for-object","title":"SRN: Side-output Residual Network for Object Reflection Symmetry Detection and Beyond","arxiv_id":"1807.06621","date":"2018-07-17","proceeding":null,"authors":["Wei Ke","Jie Chen","Jianbin Jiao","Guoying Zhao","Qixiang Ye"],"abstract":"In this paper, we establish a baseline for object reflection symmetry\ndetection in complex backgrounds by presenting a new benchmark and an\nend-to-end deep learning approach, opening up a promising direction for\nsymmetry detection in the wild. The new benchmark, Sym-PASCAL, spans challenges\nincluding object diversity, multi-objects, part-invisibility, and various\ncomplex backgrounds that are far beyond those in existing datasets. The\nend-to-end deep learning approach, referred to as a side-output residual\nnetwork (SRN), leverages the output residual units (RUs) to fit the errors\nbetween the object ground-truth symmetry and the side-outputs of multiple\nstages. By cascading RUs in a deep-to-shallow manner, SRN exploits the 'flow'\nof errors among multiple stages to address the challenges of fitting complex\noutput with limited convolutional layers, suppressing the complex backgrounds,\nand effectively matching object symmetry at different scales. SRN is further\nupgraded to a multi-task side-output residual network (MT-SRN) for joint\nsymmetry and edge detection, demonstrating its generality to image-to-mask\nlearning tasks. Experimental results validate both the challenging aspects of\nSym-PASCAL benchmark related to real-world images and the state-of-the-art\nperformance of the proposed SRN approach.","url_abs":"http://arxiv.org/abs/1807.06621v2","url_pdf":"http://arxiv.org/pdf/1807.06621v2.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":"srn-side-output-residual-network-for-object","repo_url":"https://github.com/KevinKecc/SRN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"caffe2","reach":null}],"tasks":[{"task_slug":"edge-detection","task_name":"Edge Detection"},{"task_slug":"hand-pose-estimation","task_name":"Hand Pose Estimation"},{"task_slug":"object","task_name":"Object"},{"task_slug":"symmetry-detection","task_name":"Symmetry Detection"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}