{"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/multi-scale-patch-based-representation","title":"Multi-Scale Patch-Based Representation Learning for Image Anomaly Detection and Segmentation","arxiv_id":null,"date":"2022-10-23","proceeding":"CVPR 2022 10","authors":["Chin-Chia Tsai","Tsung-Hsuan Wu","and Shang-Hong Lai"],"abstract":"Unsupervised representation learning has been proven\r\nto be effective for the challenging anomaly detection and\r\nsegmentation tasks. In this paper, we propose a multi-scale\r\npatch-based representation learning method to extract critical and representative information from normal images.\r\nBy taking the relative feature similarity between patches\r\nof different local distances into account, we can achieve\r\nbetter representation learning. Moreover, we propose a\r\nrefined way to improve the self-supervised learning strategy, thus allowing our model to learn better geometric relationship between neighboring patches. Through sliding\r\npatches of different scales all over an image, our model\r\nextracts representative features from each patch and compares them with those in the training set of normal images\r\nto detect the anomalous regions. Our experimental results\r\non MVTec AD dataset and BTAD dataset demonstrate the\r\nproposed method achieves the state-of-the-art accuracy for\r\nboth anomaly detection and segmentation.","url_abs":"https://ieeexplore.ieee.org/document/9707044","url_pdf":"https://openaccess.thecvf.com/content/WACV2022/papers/Tsai_Multi-Scale_Patch-Based_Representation_Learning_for_Image_Anomaly_Detection_and_Segmentation_WACV_2022_paper.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":[],"tasks":[{"task_slug":"anomaly-detection","task_name":"Anomaly Detection"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"self-supervised-learning","task_name":"Self-Supervised Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/anomaly-detection-on-mvtec-ad","task":"Anomaly Detection","dataset":"MVTec AD","model":"MSPBA","rank_in_archive_order":66,"of":148,"metrics":{"Detection AUROC":"98.1","Segmentation AUPRO":"95.5","Segmentation AUROC":"98.1"},"uses_additional_data":false}],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}