{"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/hypercorrelation-squeeze-for-few-shot","title":"Hypercorrelation Squeeze for Few-Shot Segmentation","arxiv_id":"2104.01538","date":"2021-04-04","proceeding":null,"authors":["Juhong Min","Dahyun Kang","Minsu Cho"],"abstract":"Few-shot semantic segmentation aims at learning to segment a target object from a query image using only a few annotated support images of the target class. This challenging task requires to understand diverse levels of visual cues and analyze fine-grained correspondence relations between the query and the support images. To address the problem, we propose Hypercorrelation Squeeze Networks (HSNet) that leverages multi-level feature correlation and efficient 4D convolutions. It extracts diverse features from different levels of intermediate convolutional layers and constructs a collection of 4D correlation tensors, i.e., hypercorrelations. Using efficient center-pivot 4D convolutions in a pyramidal architecture, the method gradually squeezes high-level semantic and low-level geometric cues of the hypercorrelation into precise segmentation masks in coarse-to-fine manner. The significant performance improvements on standard few-shot segmentation benchmarks of PASCAL-5i, COCO-20i, and FSS-1000 verify the efficacy of the proposed method.","url_abs":"https://arxiv.org/abs/2104.01538v3","url_pdf":"https://arxiv.org/pdf/2104.01538v3.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":"hypercorrelation-squeeze-for-few-shot","repo_url":"https://github.com/juhongm999/hsnet","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"feature-correlation","task_name":"Feature Correlation"},{"task_slug":"few-shot-image-segmentation","task_name":"Few-Shot Semantic Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[{"method_slug":"cp-conv","method_name":"CP conv"}],"datasets_introduced":[],"methods_introduced":[{"slug":"cp-conv","name":"CP conv","full_name":"Center-pivot convolution"}],"results":[{"leaderboard":"/sota/few-shot-semantic-segmentation-on-coco-20i-1","task":"Few-Shot Semantic Segmentation","dataset":"COCO-20i (1-shot)","model":"HSNet (ResNet-101)","rank_in_archive_order":60,"of":85,"metrics":{"FB-IoU":"69.1","Mean IoU":"41.2","learnable parameters (million)":"2.5"},"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":"HSNet (ResNet-50)","rank_in_archive_order":64,"of":85,"metrics":{"FB-IoU":"68.2","Mean IoU":"39.2","learnable parameters (million)":"2.5"},"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 (ResNet-101)","rank_in_archive_order":40,"of":81,"metrics":{"FB-IoU":"72.4","Mean IoU":"49.5","learnable parameters (million)":"2.5"},"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 (ResNet-50)","rank_in_archive_order":56,"of":81,"metrics":{"FB-IoU":"70.7","Mean IoU":"46.9","learnable parameters (million)":"2.5"},"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 (ResNet-101)","rank_in_archive_order":17,"of":24,"metrics":{"Mean IoU":"86.5"},"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 (ResNet-50)","rank_in_archive_order":21,"of":24,"metrics":{"Mean IoU":"85.5"},"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 (VGG-16)","rank_in_archive_order":23,"of":24,"metrics":{"Mean IoU":"82.3"},"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 (ResNet-101)","rank_in_archive_order":15,"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 (ResNet-50)","rank_in_archive_order":18,"of":22,"metrics":{"Mean IoU":"87.8"},"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 (VGG-16)","rank_in_archive_order":20,"of":22,"metrics":{"Mean IoU":"85.8"},"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 (ResNet-101)","rank_in_archive_order":39,"of":105,"metrics":{"FB-IoU":"77.6","Mean IoU":"66.2","learnable parameters (million)":"2.5"},"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 (ResNet-50)","rank_in_archive_order":63,"of":105,"metrics":{"FB-IoU":"76.7","Mean IoU":"64.0","learnable parameters (million)":"2.5"},"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 (VGG-16)","rank_in_archive_order":83,"of":105,"metrics":{"FB-IoU":"73.4","Mean IoU":"59.7"},"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 (ResNet-101)","rank_in_archive_order":37,"of":96,"metrics":{"FB-IoU":"80.6","Mean IoU":"70.4","learnable parameters (million)":"2.5"},"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 (ResNet-50)","rank_in_archive_order":45,"of":96,"metrics":{"FB-IoU":"80.6","Mean IoU":"69.5","learnable parameters (million)":"2.5"},"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 (VGG-16)","rank_in_archive_order":77,"of":96,"metrics":{"FB-IoU":"76.6","Mean IoU":"64.1"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2104.01538","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}