{"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/hmfs-hybrid-masking-for-few-shot-segmentation","title":"HM: Hybrid Masking for Few-Shot Segmentation","arxiv_id":"2203.12826","date":"2022-03-24","proceeding":null,"authors":["Seonghyeon Moon","Samuel S. Sohn","Honglu Zhou","Sejong Yoon","Vladimir Pavlovic","Muhammad Haris Khan","Mubbasir Kapadia"],"abstract":"We study few-shot semantic segmentation that aims to segment a target object from a query image when provided with a few annotated support images of the target class. Several recent methods resort to a feature masking (FM) technique to discard irrelevant feature activations which eventually facilitates the reliable prediction of segmentation mask. A fundamental limitation of FM is the inability to preserve the fine-grained spatial details that affect the accuracy of segmentation mask, especially for small target objects. In this paper, we develop a simple, effective, and efficient approach to enhance feature masking (FM). We dub the enhanced FM as hybrid masking (HM). Specifically, we compensate for the loss of fine-grained spatial details in FM technique by investigating and leveraging a complementary basic input masking method. Experiments have been conducted on three publicly available benchmarks with strong few-shot segmentation (FSS) baselines. We empirically show improved performance against the current state-of-the-art methods by visible margins across different benchmarks. Our code and trained models are available at: https://github.com/moonsh/HM-Hybrid-Masking","url_abs":"https://arxiv.org/abs/2203.12826v2","url_pdf":"https://arxiv.org/pdf/2203.12826v2.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":"hmfs-hybrid-masking-for-few-shot-segmentation","repo_url":"https://github.com/moonsh/hm-hybrid-masking","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"few-shot-image-segmentation","task_name":"Few-Shot Semantic Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":24,"of":85,"metrics":{"FB-IoU":"71.5","Mean IoU":"46.5"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"ASNet (HM, ResNet-101)","rank_in_archive_order":30,"of":85,"metrics":{"FB-IoU":"71.1","Mean IoU":"45.9"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"ASNet (HM, ResNet-50)","rank_in_archive_order":36,"of":85,"metrics":{"FB-IoU":"70.4","Mean IoU":"44.7"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":37,"of":85,"metrics":{"FB-IoU":"70.8","Mean IoU":"44.3"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":44,"of":85,"metrics":{"FB-IoU":"70","Mean IoU":"43.2"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-5","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (5-shot)","model":"ASNet (HM, ResNet-101)","rank_in_archive_order":35,"of":81,"metrics":{"FB-IoU":"73.3","Mean IoU":"50.6"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-5","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (5-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":36,"of":81,"metrics":{"FB-IoU":"72.9","Mean IoU":"50.6"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-5","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (5-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":43,"of":81,"metrics":{"FB-IoU":"72.2","Mean IoU":"49.4"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-5","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (5-shot)","model":"ASNet (HM, ResNet-50)","rank_in_archive_order":46,"of":81,"metrics":{"FB-IoU":"72.2","Mean IoU":"48.4"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-5","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (5-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":47,"of":81,"metrics":{"FB-IoU":"71.8","Mean IoU":"48.3"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (1-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":8,"of":13,"metrics":{"Mean IoU":"66.5"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (1-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":9,"of":13,"metrics":{"Mean IoU":"65.2"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (1-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":10,"of":13,"metrics":{"Mean IoU":"65.1"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-2","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (5-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":7,"of":12,"metrics":{"Mean IoU":"70.9"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-2","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (5-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":8,"of":12,"metrics":{"Mean IoU":"69.7"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-2","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i -> Pascal VOC (5-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":9,"of":12,"metrics":{"Mean IoU":"69.7"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-1","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (1-shot)","model":"VAT (HM, ResNet-101)","rank_in_archive_order":5,"of":24,"metrics":{"Mean IoU":"90.2"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-1","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (1-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":9,"of":24,"metrics":{"Mean IoU":"89.4"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-1","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (1-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":14,"of":24,"metrics":{"Mean IoU":"87.8"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-1","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (1-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":16,"of":24,"metrics":{"Mean IoU":"87.1"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-5","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (5-shot)","model":"VAT (HM, ResNet-101)","rank_in_archive_order":6,"of":22,"metrics":{"Mean IoU":"90.5"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-5","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (5-shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":8,"of":22,"metrics":{"Mean IoU":"89.9"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-5","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (5-shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":14,"of":22,"metrics":{"Mean IoU":"88.5"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-fss-1000-5","task":"Few-Shot Semantic Segmentation","dataset":"FSS-1000 (5-shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":16,"of":22,"metrics":{"Mean IoU":"88"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-1","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (1-Shot)","model":"VAT (HM, ResNet-101)","rank_in_archive_order":23,"of":105,"metrics":{"FB-IoU":"79.4","Mean IoU":"67.8"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-1","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (1-Shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":36,"of":105,"metrics":{"FB-IoU":"77.8","Mean IoU":"66.7"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-1","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (1-Shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":43,"of":105,"metrics":{"FB-IoU":"77.1","Mean IoU":"65.8"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-1","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (1-Shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":51,"of":105,"metrics":{"FB-IoU":"76.5","Mean IoU":"65"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-5","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (5-Shot)","model":"VAT (HM, ResNet-101)","rank_in_archive_order":32,"of":96,"metrics":{"FB-IoU":"81.5","Mean IoU":"70.9"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-5","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (5-Shot)","model":"HSNet (HM, ResNet-101)","rank_in_archive_order":47,"of":96,"metrics":{"FB-IoU":"79.7","Mean IoU":"69.3"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-5","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (5-Shot)","model":"VAT (HM, ResNet-50)","rank_in_archive_order":59,"of":96,"metrics":{"FB-IoU":"78.5","Mean IoU":"68.2"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-semantic-segmentation-on-pascal-5i-5","task":"Few-Shot Semantic Segmentation","dataset":"PASCAL-5i (5-Shot)","model":"HSNet (HM, ResNet-50)","rank_in_archive_order":63,"of":96,"metrics":{"FB-IoU":"77.7","Mean IoU":"67.1"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2203.12826","atlas_url":"https://app.syntology.ai/?focus=2203.12826","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2203.12826"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/moonsh/HM-Hybrid-Masking","reach":null}],"summary":{"unverified":1},"by_repo_kind":{"official":{"samples":1,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":1,"samples":[{"code_sha256_prefix":"3af45cd405738bcc","entry":"HypercorrSqueezeNetwork","repo":"moonsh/HM-Hybrid-Masking","repo_kind":"official","path":"hsnet.py","file_url":"https://github.com/moonsh/HM-Hybrid-Masking/blob/HEAD/hsnet.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3af45cd405738bcc"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}