{"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/190407220","title":"Learning Discriminative Model Prediction for Tracking","arxiv_id":"1904.07220","date":"2019-04-15","proceeding":"ICCV 2019 10","authors":["Goutam Bhat","Martin Danelljan","Luc van Gool","Radu Timofte"],"abstract":"The current strive towards end-to-end trainable computer vision systems imposes major challenges for the task of visual tracking. In contrast to most other vision problems, tracking requires the learning of a robust target-specific appearance model online, during the inference stage. To be end-to-end trainable, the online learning of the target model thus needs to be embedded in the tracking architecture itself. Due to the imposed challenges, the popular Siamese paradigm simply predicts a target feature template, while ignoring the background appearance information during inference. Consequently, the predicted model possesses limited target-background discriminability. We develop an end-to-end tracking architecture, capable of fully exploiting both target and background appearance information for target model prediction. Our architecture is derived from a discriminative learning loss by designing a dedicated optimization process that is capable of predicting a powerful model in only a few iterations. Furthermore, our approach is able to learn key aspects of the discriminative loss itself. The proposed tracker sets a new state-of-the-art on 6 tracking benchmarks, achieving an EAO score of 0.440 on VOT2018, while running at over 40 FPS. The code and models are available at https://github.com/visionml/pytracking.","url_abs":"https://arxiv.org/abs/1904.07220v2","url_pdf":"https://arxiv.org/pdf/1904.07220v2.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":"190407220","repo_url":"https://github.com/visionml/pytracking","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"190407220","repo_url":"https://github.com/martin-danelljan/ECO","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[{"task_slug":"object-tracking","task_name":"Object Tracking"},{"task_slug":"prediction","task_name":"Prediction"},{"task_slug":"video-object-tracking","task_name":"Video Object Tracking"},{"task_slug":"visual-object-tracking","task_name":"Visual Object Tracking"},{"task_slug":"visual-tracking","task_name":"Visual Tracking"},{"task_slug":"model","task_name":"model"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/object-tracking-on-fe108","task":"Object Tracking","dataset":"FE108","model":"DiMP","rank_in_archive_order":5,"of":8,"metrics":{"Averaged Precision":"85.1","Success Rate":"57.1"},"uses_additional_data":false},{"leaderboard":"/sota/video-object-tracking-on-nv-vot211","task":"Video Object Tracking","dataset":"NT-VOT211","model":"DiMP-50","rank_in_archive_order":18,"of":43,"metrics":{"AUC":"35.89","Precision":"48.68"},"uses_additional_data":false},{"leaderboard":"/sota/visual-object-tracking-on-got-10k","task":"Visual Object Tracking","dataset":"GOT-10k","model":"DiMP","rank_in_archive_order":40,"of":42,"metrics":{"Average Overlap":"61.1","Success Rate 0.5":"71.7"},"uses_additional_data":false},{"leaderboard":"/sota/visual-object-tracking-on-lasot","task":"Visual Object Tracking","dataset":"LaSOT","model":"DiMP","rank_in_archive_order":43,"of":46,"metrics":{"AUC":"56.8","Normalized Precision":"65.0","Precision":"56.7"},"uses_additional_data":false},{"leaderboard":"/sota/visual-object-tracking-on-lasot","task":"Visual Object Tracking","dataset":"LaSOT","model":"DiMP-50","rank_in_archive_order":45,"of":46,"metrics":{"Precision":"68.7"},"uses_additional_data":false},{"leaderboard":"/sota/visual-object-tracking-on-trackingnet","task":"Visual Object Tracking","dataset":"TrackingNet","model":"DiMP-50","rank_in_archive_order":32,"of":40,"metrics":{"Accuracy":"74.0","Normalized Precision":"80.1"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1904.07220","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}