{"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/mst-adaptive-multi-scale-tokens-guided","title":"MST: Adaptive Multi-Scale Tokens Guided Interactive Segmentation","arxiv_id":"2401.04403","date":"2024-01-09","proceeding":null,"authors":["Long Xu","Shanghong Li","Yongquan Chen","Jun Luo","Shiwu Lai"],"abstract":"Interactive segmentation has gained significant attention for its application in human-computer interaction and data annotation. To address the target scale variation issue in interactive segmentation, a novel multi-scale token adaptation algorithm is proposed. By performing top-k operations across multi-scale tokens, the computational complexity is greatly simplified while ensuring performance. To enhance the robustness of multi-scale token selection, we also propose a token learning algorithm based on contrastive loss. This algorithm can effectively improve the performance of multi-scale token adaptation. Extensive benchmarking shows that the algorithm achieves state-of-the-art (SOTA) performance, compared to current methods. An interactive demo and all reproducible codes will be released at https://github.com/hahamyt/mst.","url_abs":"https://arxiv.org/abs/2401.04403v2","url_pdf":"https://arxiv.org/pdf/2401.04403v2.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":"mst-adaptive-multi-scale-tokens-guided","repo_url":"https://github.com/hahamyt/mst","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"benchmarking","task_name":"Benchmarking"},{"task_slug":"interactive-segmentation","task_name":"Interactive Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/interactive-segmentation-on-berkeley","task":"Interactive Segmentation","dataset":"Berkeley","model":"ViT-B+MST+CL","rank_in_archive_order":3,"of":14,"metrics":{"NoC@90":"1.50"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-coco-minival","task":"Interactive Segmentation","dataset":"COCO minival","model":"ViT-B+MST+CL","rank_in_archive_order":1,"of":1,"metrics":{"NoC@85":"2.08","NoC@90":"2.85"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-davis","task":"Interactive Segmentation","dataset":"DAVIS","model":"ViT-B+MST+CL","rank_in_archive_order":3,"of":15,"metrics":{"NoC@90":"4.55"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-davis-585","task":"Interactive Segmentation","dataset":"DAVIS-585","model":"ViT-B+MST+CL","rank_in_archive_order":1,"of":2,"metrics":{"NoC@85":"1.80","NoC@90":"2.29"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-grabcut","task":"Interactive Segmentation","dataset":"GrabCut","model":"ViT-B+MST+CL","rank_in_archive_order":5,"of":18,"metrics":{"NoC@90":"1.48"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-pascalvoc","task":"Interactive Segmentation","dataset":"PascalVOC","model":"ViT-B+MST+CL","rank_in_archive_order":1,"of":1,"metrics":{"NoC@85":"1.69","NoC@90":"1.90"},"uses_additional_data":false},{"leaderboard":"/sota/interactive-segmentation-on-sbd","task":"Interactive Segmentation","dataset":"SBD","model":"ViT-B+MST+CL","rank_in_archive_order":13,"of":14,"metrics":{"NoC@85":"3.03"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}