{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T17:51:31Z","timestamp":1791827491016,"version":"4.3.4"},"reference-count":40,"publisher":"Public Library of Science (PLoS)","issue":"10","license":[{"start":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T00:00:00Z","timestamp":1791763200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Temporal Action Localization (TAL) aims to identify all action categories along with their precise temporal boundaries in untrimmed videos. It serves as a fundamental task supporting various downstream applications, including intelligent video editing, automatic summarization, anomaly detection, and human\u2013computer interaction. Despite significant progress, TAL in real-world scenarios remains challenging due to two critical factors: substantial intra-class variations in action duration, ranging from a few seconds to several minutes, and severe degradation of boundary localization accuracy caused by variations in illumination, background clutter, and camera viewpoints. To address these issues, this paper presents a Temporal Action Localization with Class-Guided Boundary Regression (TAL-CGBR) framework, which incorporates semantic information from the classification branch into the regression network to enhance boundary precision. In addition, a window-based attention mechanism and a temporal hierarchical feature structure are introduced to improve computational efficiency while maintaining high localization accuracy. Experimental evaluations on benchmark datasets demonstrate that the proposed TAL-CGBR achieves competitive performance, attaining 68.5% average mAP on THUMOS14 and 36.7% average mAP on ActivityNet v1.3, ranking second among recent methods and demonstrating a favorable balance between accuracy and computational efficiency.<\/jats:p>","DOI":"10.1371\/journal.pone.0360160","type":"journal-article","created":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T17:29:29Z","timestamp":1791826169000},"page":"e0360160","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["TAL-CGBR: A class-guided boundary regression approach for Temporal Action Localization"],"prefix":"10.1371","volume":"21","author":[{"given":"Jinling","family":"Wei","sequence":"first","affiliation":[{"name":"Hangzhou City University","place":["Hangzhou, China"],"department":["School of Computer and Computing Science"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haolong","family":"Chai","sequence":"additional","affiliation":[{"name":"Hangzhou City University","place":["Hangzhou, China"],"department":["School of Computer and Computing Science"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wenwu","family":"Hou","sequence":"additional","affiliation":[{"name":"Zhejiang University","place":["Hangzhou, China"],"department":["College of Computer Science and Technology"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-0457-2505","authenticated-orcid":true,"given":"Hongwen","family":"Yuan","sequence":"additional","affiliation":[{"name":"Hangzhou City University","place":["Hangzhou, China"],"department":["Department of Physical Education and Aesthetic Education"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiao","family":"Huang","sequence":"additional","affiliation":[{"name":"Hangzhou City University","place":["Hangzhou, China"],"department":["Department of Physical Education and Aesthetic Education"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sufen","family":"Huang","sequence":"additional","affiliation":[{"name":"Hangzhou City University","place":["Hangzhou, China"],"department":["Department of Physical Education and Aesthetic Education"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,10,12]]},"reference":[{"key":"pone.0360160.ref001","first-page":"1130","author":"Y-W Chao","year":"2018"},{"issue":"2","key":"pone.0360160.ref002","article-title":"Overview of temporal action detection based on deep learning","volume":"57","author":"K Hu","year":"2024","journal-title":"Artif Intell Rev"},{"key":"pone.0360160.ref003","author":"C Han","year":"2026"},{"key":"pone.0360160.ref004","doi-asserted-by":"crossref","unstructured":"Zhang CL, Wu J, Li Y. \u201cActionformer: Localizing moments of actions with transformers,\u201d in European Conference on Computer Vision. 2022, pp. 492\u2013510.","DOI":"10.1007\/978-3-031-19772-7_29"},{"key":"pone.0360160.ref005","doi-asserted-by":"crossref","first-page":"5427","DOI":"10.1109\/TIP.2022.3195321","article-title":"End-to-End Temporal Action Detection With Transformer","volume":"31","author":"X Liu","year":"2022","journal-title":"IEEE Trans Image Process"},{"key":"pone.0360160.ref006","doi-asserted-by":"crossref","unstructured":"Cheng F, Bertasius G. \u201cTallformer: Temporal action localization with a long-memory transformer,\u201d in European Conference on Computer Vision. 2022; pp. 503\u201321.","DOI":"10.1007\/978-3-031-19830-4_29"},{"key":"pone.0360160.ref007","doi-asserted-by":"crossref","unstructured":"Shi D, Zhong Y, Cao Q, Zhang J, Ma L, Li J, et al. \u201cReact: Temporal action detection with relational queries,\u201d in European conference on computer vision. 2022; pp. 105\u201321.","DOI":"10.1007\/978-3-031-20080-9_7"},{"key":"pone.0360160.ref008","first-page":"13496","author":"Z Zhu","year":"2021"},{"key":"pone.0360160.ref009","first-page":"13411","author":"J Shao","year":"2023"},{"key":"pone.0360160.ref010","first-page":"10328","author":"S Nag","year":"2023"},{"key":"pone.0360160.ref011","first-page":"13638","author":"C Zhao","year":"2021"},{"issue":"4","key":"pone.0360160.ref012","doi-asserted-by":"crossref","first-page":"2171","DOI":"10.1109\/TPAMI.2023.3330794","article-title":"Temporal Action Localization in the Deep Learning Era: A Survey","volume":"46","author":"B Wang","year":"2024","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"pone.0360160.ref013","author":"L Wang","year":"2021"},{"key":"pone.0360160.ref014","first-page":"13506","author":"J Tan","year":"2021"},{"key":"pone.0360160.ref015","author":"X Zhu","year":"2020"},{"issue":"3","key":"pone.0360160.ref016","doi-asserted-by":"crossref","first-page":"3644","DOI":"10.1609\/aaai.v36i3.20277","article-title":"Learning Disentangled Classification and Localization Representations for Temporal Action Localization","volume":"36","author":"Z Zhu","year":"2022","journal-title":"AAAI"},{"key":"pone.0360160.ref017","doi-asserted-by":"crossref","unstructured":"Shao J, Wang X, Quan R, Zheng J, Yang J, Yang Y. Action sensitivity learning for temporal action localization. In: 2023 IEEE\/CVF International Conference on Computer Vision (ICCV). 2023;13411\u201323.","DOI":"10.1109\/ICCV51070.2023.01238"},{"key":"pone.0360160.ref018","first-page":"7093","author":"R Zeng","year":"2019"},{"issue":"10","key":"pone.0360160.ref019","doi-asserted-by":"crossref","first-page":"6209","DOI":"10.1109\/TPAMI.2021.3090167","article-title":"Graph convolutional module for temporal action localization in videos","volume":"44","author":"R Zeng","year":"2022","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"pone.0360160.ref020","article-title":"Enhancing the locality and breaking the memory bottleneck of transformer on time series forecasting","volume":"32","author":"S Li","year":"2019","journal-title":"Advances in neural information processing systems"},{"key":"pone.0360160.ref021","first-page":"30392","article-title":"Early convolutions help transformers see better","volume":"34","author":"T Xiao","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"pone.0360160.ref022","unstructured":"Choromanski K, Likhosherstov V, Dohan D, Song X, Gane A, Sarlos T, et al. \u201cRethinking attention with performers,\u201d arXiv preprint arXiv. 2020."},{"issue":"16","key":"pone.0360160.ref023","first-page":"14138","article-title":"Nystr\u00f6mformer: A Nyst\u00f6m-based Algorithm for Approximating Self-Attention","volume":"35","author":"Y Xiong","year":"2021","journal-title":"Proc AAAI Conf Artif Intell"},{"key":"pone.0360160.ref024","article-title":"Attention is all you need","volume":"30","author":"A Vaswani","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"pone.0360160.ref025","author":"TN Tang","year":"2023"},{"key":"pone.0360160.ref026","first-page":"3888","author":"T Lin","year":"2019"},{"issue":"07","key":"pone.0360160.ref027","doi-asserted-by":"crossref","first-page":"12993","DOI":"10.1609\/aaai.v34i07.6999","article-title":"Distance-IoU Loss: Faster and Better Learning for Bounding Box Regression","volume":"34","author":"Z Zheng","year":"2020","journal-title":"AAAI"},{"key":"pone.0360160.ref028","first-page":"4724","author":"J Carreira","year":"2017"},{"key":"pone.0360160.ref029","author":"W Kay","year":"2017"},{"key":"pone.0360160.ref030","article-title":"Pytorch: An imperative style, high-performance deep learning library","volume":"32","author":"A Paszke","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"pone.0360160.ref031","author":"I Loshchilov","year":"2017"},{"key":"pone.0360160.ref032","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.cviu.2016.10.018","article-title":"The THUMOS challenge on action recognition for videos \u201cin the wild\u201d","volume":"155","author":"H Idrees","year":"2017","journal-title":"Computer Vision and Image Understanding"},{"key":"pone.0360160.ref033","first-page":"3319","author":"C Lin","year":"2021"},{"key":"pone.0360160.ref034","first-page":"961","author":"FC Heilbron","year":"2015"},{"key":"pone.0360160.ref035","first-page":"18591","author":"S Liu","year":"2024"},{"key":"pone.0360160.ref036","first-page":"6201","author":"C Feichtenhofer","year":"2019"},{"issue":"5","key":"pone.0360160.ref037","doi-asserted-by":"crossref","first-page":"4878","DOI":"10.1609\/aaai.v39i5.32516","article-title":"Temporal action localization with cross layer task decoupling and refinement","volume":"39","author":"Q Li","year":"2025","journal-title":"AAAI"},{"key":"pone.0360160.ref038","first-page":"112144","article-title":"Long-term pre-training for temporal action detection with transformers","author":"J Kim","year":"2025","journal-title":"Pattern Recognition"},{"key":"pone.0360160.ref039","doi-asserted-by":"crossref","unstructured":"Nag S, Zhu X, Song YZ, Xiang T. Zero-Shot Temporal Action Detection via Vision-Language Prompting. In: Computer Vision \u2013 ECCV 2022. 2022;681\u201397.","DOI":"10.1007\/978-3-031-20062-5_39"},{"key":"pone.0360160.ref040","first-page":"19978","author":"X Liu","year":"2022"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0360160","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,12]],"date-time":"2026-10-12T17:29:34Z","timestamp":1791826174000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0360160"}},"subtitle":[],"editor":[{"given":"Xian-Shi","family":"Zhang","sequence":"first","affiliation":[{"name":"University of Electronic Science and Technology of China","place":["CHINA"]}],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,10,12]]},"references-count":40,"journal-issue":{"issue":"10","published-online":{"date-parts":[[2026,10,12]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0360160","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,12]]}}}