{"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/mobilevos-real-time-video-object-segmentation","title":"MobileVOS: Real-Time Video Object Segmentation Contrastive Learning meets Knowledge Distillation","arxiv_id":"2303.07815","date":"2023-03-14","proceeding":"CVPR 2023 1","authors":["Roy Miles","Mehmet Kerim Yucel","Bruno Manganelli","Albert Saa-Garriga"],"abstract":"This paper tackles the problem of semi-supervised video object segmentation on resource-constrained devices, such as mobile phones. We formulate this problem as a distillation task, whereby we demonstrate that small space-time-memory networks with finite memory can achieve competitive results with state of the art, but at a fraction of the computational cost (32 milliseconds per frame on a Samsung Galaxy S22). Specifically, we provide a theoretically grounded framework that unifies knowledge distillation with supervised contrastive representation learning. These models are able to jointly benefit from both pixel-wise contrastive learning and distillation from a pre-trained teacher. We validate this loss by achieving competitive J&F to state of the art on both the standard DAVIS and YouTube benchmarks, despite running up to 5x faster, and with 32x fewer parameters.","url_abs":"https://arxiv.org/abs/2303.07815v1","url_pdf":"https://arxiv.org/pdf/2303.07815v1.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":"contrastive-learning","task_name":"Contrastive Learning"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"semi-supervised-video-object-segmentation","task_name":"Semi-Supervised Video Object Segmentation"},{"task_slug":"video-object-segmentation","task_name":"Video Object Segmentation"},{"task_slug":"video-semantic-segmentation","task_name":"Video Semantic Segmentation"}],"methods":[{"method_slug":"contrastive-learning","method_name":"Contrastive Learning"},{"method_slug":"knowledge-distillation","method_name":"Knowledge Distillation"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/visual-object-tracking-on-davis-2016","task":"Semi-Supervised Video Object Segmentation","dataset":"DAVIS 2016","model":"MobileVOS (BL30K)","rank_in_archive_order":20,"of":78,"metrics":{"F-measure (Mean)":"92.6","J&F":"91.4","Jaccard (Mean)":"90.3","Speed (FPS)":"100.1"},"uses_additional_data":true},{"leaderboard":"/sota/visual-object-tracking-on-davis-2016","task":"Semi-Supervised Video Object Segmentation","dataset":"DAVIS 2016","model":"MobileVOS","rank_in_archive_order":26,"of":78,"metrics":{"F-measure (Mean)":"91.6","J&F":"90.6","Jaccard (Mean)":"89.7","Speed (FPS)":"100.1"},"uses_additional_data":false},{"leaderboard":"/sota/visual-object-tracking-on-davis-2017","task":"Semi-Supervised Video Object Segmentation","dataset":"DAVIS 2017 (val)","model":"MobileVOS (BL30K)","rank_in_archive_order":38,"of":81,"metrics":{"F-measure (Mean)":"88.9","J&F":"82.3","Params(M)":"8.1","Speed (FPS)":"90.6"},"uses_additional_data":true},{"leaderboard":"/sota/visual-object-tracking-on-davis-2017","task":"Semi-Supervised Video Object Segmentation","dataset":"DAVIS 2017 (val)","model":"MobileVOS","rank_in_archive_order":45,"of":81,"metrics":{"F-measure (Mean)":"87.1","J&F":"80.2","Params(M)":"8.1","Speed (FPS)":"90.6"},"uses_additional_data":false},{"leaderboard":"/sota/video-object-segmentation-on-davis-2016","task":"Video Object Segmentation","dataset":"DAVIS 2016","model":"MobileVOS (val)","rank_in_archive_order":6,"of":24,"metrics":{"F-Score":"92.6","J&F":"91.4","Jaccard (Mean)":"90.3"},"uses_additional_data":false},{"leaderboard":"/sota/video-object-segmentation-on-youtube-vos-2019-2","task":"Video Object Segmentation","dataset":"YouTube-VOS 2019","model":"MobileVOS","rank_in_archive_order":6,"of":10,"metrics":{"F-Measure (Seen)":"87.7","F-Measure (Unseen)":"85.3","Jaccard (Seen)":"83.2","Jaccard (Unseen)":"76.9","Mean Jaccard & F-Measure":"83.3"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=2303.07815","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}