{"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/zero-shot-detection","title":"Zero-Shot Detection","arxiv_id":"1803.07113","date":"2018-03-19","proceeding":null,"authors":["Pengkai Zhu","Hanxiao Wang","Venkatesh Saligrama"],"abstract":"As we move towards large-scale object detection, it is unrealistic to expect\nannotated training data, in the form of bounding box annotations around\nobjects, for all object classes at sufficient scale, and so methods capable of\nunseen object detection are required. We propose a novel zero-shot method based\non training an end-to-end model that fuses semantic attribute prediction with\nvisual features to propose object bounding boxes for seen and unseen classes.\nWhile we utilize semantic features during training, our method is agnostic to\nsemantic information for unseen classes at test-time. Our method retains the\nefficiency and effectiveness of YOLOv2 for objects seen during training, while\nimproving its performance for novel and unseen objects. The ability of\nstate-of-art detection methods to learn discriminative object features to\nreject background proposals also limits their performance for unseen objects.\nWe posit that, to detect unseen objects, we must incorporate semantic\ninformation into the visual domain so that the learned visual features reflect\nthis information and leads to improved recall rates for unseen objects. We test\nour method on PASCAL VOC and MS COCO dataset and observed significant\nimprovements on the average precision of unseen classes.","url_abs":"http://arxiv.org/abs/1803.07113v2","url_pdf":"http://arxiv.org/pdf/1803.07113v2.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":"zero-shot-detection","repo_url":"https://github.com/howBiGaStorm/ZeroShot-YOLO","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-detection","task_name":"Object Detection"},{"task_slug":"object-detection-1","task_name":"object-detection"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"darknet-19","method_name":"Darknet-19"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"yolov2","method_name":"YOLOv2"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1803.07113","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}