{"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/lstd-a-low-shot-transfer-detector-for-object","title":"LSTD: A Low-Shot Transfer Detector for Object Detection","arxiv_id":"1803.01529","date":"2018-03-05","proceeding":null,"authors":["Hao Chen","Yali Wang","Guoyou Wang","Yu Qiao"],"abstract":"Recent advances in object detection are mainly driven by deep learning with\nlarge-scale detection benchmarks. However, the fully-annotated training set is\noften limited for a target detection task, which may deteriorate the\nperformance of deep detectors. To address this challenge, we propose a novel\nlow-shot transfer detector (LSTD) in this paper, where we leverage rich\nsource-domain knowledge to construct an effective target-domain detector with\nvery few training examples. The main contributions are described as follows.\nFirst, we design a flexible deep architecture of LSTD to alleviate transfer\ndifficulties in low-shot detection. This architecture can integrate the\nadvantages of both SSD and Faster RCNN in a unified deep framework. Second, we\nintroduce a novel regularized transfer learning framework for low-shot\ndetection, where the transfer knowledge (TK) and background depression (BD)\nregularizations are proposed to leverage object knowledge respectively from\nsource and target domains, in order to further enhance fine-tuning with a few\ntarget images. Finally, we examine our LSTD on a number of challenging low-shot\ndetection experiments, where LSTD outperforms other state-of-the-art\napproaches. The results demonstrate that LSTD is a preferable deep detector for\nlow-shot scenarios.","url_abs":"http://arxiv.org/abs/1803.01529v1","url_pdf":"http://arxiv.org/pdf/1803.01529v1.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":"lstd-a-low-shot-transfer-detector-for-object","repo_url":"https://github.com/Cassie94/LSTD","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"few-shot-object-detection","task_name":"Few-Shot Object Detection"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"non-maximum-suppression","method_name":"Non Maximum Suppression"},{"method_slug":"ssd","method_name":"SSD"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/few-shot-object-detection-on-ms-coco-10-shot","task":"Few-Shot Object Detection","dataset":"MS-COCO (10-shot)","model":"LSTD (YOLO)","rank_in_archive_order":33,"of":33,"metrics":{"AP":"3.2"},"uses_additional_data":false},{"leaderboard":"/sota/few-shot-object-detection-on-ms-coco-30-shot","task":"Few-Shot Object Detection","dataset":"MS-COCO (30-shot)","model":"LSTD (YOLO)","rank_in_archive_order":25,"of":25,"metrics":{"AP":"6.7"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1803.01529","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}