{"id":1674,"date":"2025-03-29T20:09:00","date_gmt":"2025-03-29T12:09:00","guid":{"rendered":"https:\/\/tongjicdi.com\/?p=1674"},"modified":"2025-05-29T21:13:27","modified_gmt":"2025-05-29T13:13:27","slug":"toward-open-vocabulary-aerial-object-detection-with-clip-activated-student-teacher-learning","status":"publish","type":"post","link":"https:\/\/tongjicdi.com\/index.php\/2025\/03\/29\/toward-open-vocabulary-aerial-object-detection-with-clip-activated-student-teacher-learning\/","title":{"rendered":"Toward Open Vocabulary Aerial Object Detection with CLIP-Activated Student-Teacher Learning"},"content":{"rendered":"\t\t<div data-elementor-type=\"wp-post\" data-elementor-id=\"1674\" class=\"elementor elementor-1674\" data-elementor-post-type=\"post\">\n\t\t\t\t<div class=\"elementor-element elementor-element-55b06f2 e-flex e-con-boxed e-con e-parent\" data-id=\"55b06f2\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-3035e450 elementor-widget elementor-widget-theme-post-featured-image elementor-widget-image\" data-id=\"3035e450\" data-element_type=\"widget\" data-widget_type=\"theme-post-featured-image.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.22.0 - 24-06-2024 *\/\n.elementor-widget-image{text-align:center}.elementor-widget-image a{display:inline-block}.elementor-widget-image a img[src$=\".svg\"]{width:48px}.elementor-widget-image img{vertical-align:middle;display:inline-block}<\/style>\t\t\t\t\t\t\t\t\t\t<img fetchpriority=\"high\" decoding=\"async\" width=\"800\" height=\"606\" src=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41.png\" class=\"attachment-large size-large wp-image-1675\" alt=\"\" srcset=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41.png 856w, https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41-300x227.png 300w, https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41-768x581.png 768w\" sizes=\"(max-width: 800px) 100vw, 800px\" \/>\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-1fd0e498 elementor-widget elementor-widget-theme-post-title elementor-page-title elementor-widget-heading\" data-id=\"1fd0e498\" data-element_type=\"widget\" data-widget_type=\"theme-post-title.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.22.0 - 24-06-2024 *\/\n.elementor-heading-title{padding:0;margin:0;line-height:1}.elementor-widget-heading .elementor-heading-title[class*=elementor-size-]>a{color:inherit;font-size:inherit;line-height:inherit}.elementor-widget-heading .elementor-heading-title.elementor-size-small{font-size:15px}.elementor-widget-heading .elementor-heading-title.elementor-size-medium{font-size:19px}.elementor-widget-heading .elementor-heading-title.elementor-size-large{font-size:29px}.elementor-widget-heading .elementor-heading-title.elementor-size-xl{font-size:39px}.elementor-widget-heading .elementor-heading-title.elementor-size-xxl{font-size:59px}<\/style><h1 class=\"elementor-heading-title elementor-size-default\">Toward Open Vocabulary Aerial Object Detection with CLIP-Activated Student-Teacher Learning<\/h1>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-37380d43 elementor-widget elementor-widget-text-editor\" data-id=\"37380d43\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.22.0 - 24-06-2024 *\/\n.elementor-widget-text-editor.elementor-drop-cap-view-stacked .elementor-drop-cap{background-color:#69727d;color:#fff}.elementor-widget-text-editor.elementor-drop-cap-view-framed .elementor-drop-cap{color:#69727d;border:3px solid;background-color:transparent}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap{margin-top:8px}.elementor-widget-text-editor:not(.elementor-drop-cap-view-default) .elementor-drop-cap-letter{width:1em;height:1em}.elementor-widget-text-editor .elementor-drop-cap{float:left;text-align:center;line-height:1;font-size:50px}.elementor-widget-text-editor .elementor-drop-cap-letter{display:inline-block}<\/style>\t\t\t\t<p>Yan Li, Weiwei Guo, Xue Yang, Ning Liao, Dunyun He, Jiaqi Zhou, Wenxian Yu<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-186c5b82 elementor-widget elementor-widget-text-editor\" data-id=\"186c5b82\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>European Conference on Computer Vision (2025)<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-3a5b3a11 e-flex e-con-boxed e-con e-parent\" data-id=\"3a5b3a11\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-ce8f604 elementor-widget elementor-widget-heading\" data-id=\"ce8f604\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h2 class=\"elementor-heading-title elementor-size-default\">\u6458\u8981\/Abstract<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-2da803a3 e-flex e-con-boxed e-con e-parent\" data-id=\"2da803a3\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t<div class=\"elementor-element elementor-element-6a7e9131 e-con-full e-flex e-con e-child\" data-id=\"6a7e9131\" data-element_type=\"container\">\n\t\t\t\t<div class=\"elementor-element elementor-element-30aed0de elementor-widget elementor-widget-text-editor\" data-id=\"30aed0de\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>\u8d8a\u6765\u8d8a\u591a\u7684\u9065\u611f\u56fe\u50cf\u523a\u6fc0\u4e86\u53ef\u6269\u5c55\u5bf9\u8c61\u68c0\u6d4b\u5668\u7684\u53d1\u5c55\uff0c\u8fd9\u4e9b\u68c0\u6d4b\u5668\u53ef\u4ee5\u68c0\u6d4b\u8bad\u7ec3\u7c7b\u522b\u4e4b\u5916\u7684\u5bf9\u8c61\uff0c\u800c\u65e0\u9700\u82b1\u8d39\u5927\u91cf\u6210\u672c\u6536\u96c6\u65b0\u7684\u6807\u8bb0\u6570\u636e\u3002\u5728\u672c\u6587\u4e2d\uff0c\u6211\u4eec\u7684\u76ee\u6807\u662f\u5f00\u53d1\u822a\u62cd\u56fe\u50cf\u4e2d\u7684\u5f00\u653e\u8bcd\u6c47\u76ee\u6807\u68c0\u6d4b\uff08OVD\uff09\u6280\u672f\uff0c\u5c06\u76ee\u6807\u8bcd\u6c47\u91cf\u6269\u5927\u5230\u8bad\u7ec3\u6570\u636e\u4e4b\u5916\u3002 OVD \u7684\u6027\u80fd\u5f88\u5927\u7a0b\u5ea6\u4e0a\u4f9d\u8d56\u4e8e\u4e0e\u7c7b\u65e0\u5173\u7684\u533a\u57df\u63d0\u8bae\u548c\u65b0\u5bf9\u8c61\u7c7b\u522b\u7684\u4f2a\u6807\u7b7e\u7684\u8d28\u91cf\u3002\u4e3a\u4e86\u540c\u65f6\u751f\u6210\u9ad8\u8d28\u91cf\u7684\u63d0\u6848\u548c\u4f2a\u6807\u7b7e\uff0c\u6211\u4eec\u63d0\u51fa\u4e86CastDet \uff0c\u4e00\u4e2a\u7531CLIP\u6fc0\u6d3b\u7684\u5b66\u751f\u6559\u5e08\u5f00\u653e\u8bcd\u6c47\u5bf9\u8c61\u68c0\u6d4b\u6846\u67b6\u3002\u6211\u4eec\u7684\u7aef\u5230\u7aef\u6846\u67b6\u9075\u5faa\u5b66\u751f-\u6559\u5e08\u81ea\u5b66\u4e60\u673a\u5236\uff0c\u91c7\u7528 RemoteCLIP \u6a21\u578b\u4f5c\u4e3a\u4e00\u4f4d\u62e5\u6709\u4e30\u5bcc\u77e5\u8bc6\u7684\u989d\u5916\u5168\u77e5\u6559\u5e08\u3002\u901a\u8fc7\u8fd9\u6837\u505a\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u4e0d\u4ec5\u4fc3\u8fdb\u4e86\u65b0\u9896\u7684\u5bf9\u8c61\u63d0\u8bae\uff0c\u800c\u4e14\u4fc3\u8fdb\u4e86\u5206\u7c7b\u3002\u6b64\u5916\uff0c\u6211\u4eec\u8bbe\u8ba1\u4e86\u4e00\u79cd\u52a8\u6001\u6807\u7b7e\u961f\u5217\u7b56\u7565\uff0c\u4ee5\u5728\u6279\u91cf\u8bad\u7ec3\u671f\u95f4\u4fdd\u6301\u9ad8\u8d28\u91cf\u7684\u4f2a\u6807\u7b7e\u3002\u6211\u4eec\u5bf9\u591a\u4e2a\u73b0\u6709\u7684\u7a7a\u4e2d\u7269\u4f53\u68c0\u6d4b\u6570\u636e\u96c6\u8fdb\u884c\u4e86\u5e7f\u6cdb\u7684\u5b9e\u9a8c\uff0c\u8fd9\u4e9b\u6570\u636e\u96c6\u662f\u4e3a OVD \u4efb\u52a1\u8bbe\u7f6e\u7684\u3002\u5b9e\u9a8c\u7ed3\u679c\u8868\u660e\uff0c\u6211\u4eec\u7684 CastDet \u5b9e\u73b0\u4e86\u5353\u8d8a\u7684\u5f00\u653e\u8bcd\u6c47\u68c0\u6d4b\u6027\u80fd\uff0c\u4f8b\u5982\uff0c\u5728 VisDroneZSD \u65b0\u9896\u7c7b\u522b\u4e0a\u8fbe\u5230 46.5% mAP\uff0c\u6bd4\u6700\u5148\u8fdb\u7684\u5f00\u653e\u8bcd\u6c47\u68c0\u6d4b\u5668\u9ad8\u51fa 21.0% mAP\u3002\u636e\u6211\u4eec\u6240\u77e5\uff0c\u8fd9\u662f\u7b2c\u4e00\u4e2a\u5e94\u7528\u548c\u5f00\u53d1\u822a\u7a7a\u56fe\u50cf\u5f00\u653e\u8bcd\u6c47\u76ee\u6807\u68c0\u6d4b\u6280\u672f\u7684\u5de5\u4f5c\u3002\u8be5\u4ee3\u7801\u53ef\u5728https:\/\/github.com\/lizzy8587\/CastDet\u83b7\u53d6\u3002<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-236f6cef e-con-full e-flex e-con e-child\" data-id=\"236f6cef\" data-element_type=\"container\">\n\t\t\t\t<div class=\"elementor-element elementor-element-2d81e00d elementor-widget elementor-widget-text-editor\" data-id=\"2d81e00d\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\tAn increasingly massive number of remote-sensing images spurs the development of extensible object detectors that can detect objects beyond training categories without costly collecting new labeled data. In this paper, we aim to develop open-vocabulary object detection (OVD) technique in aerial images that scales up object vocabulary size beyond training data. The performance of OVD greatly relies on the quality of class-agnostic region proposals and pseudo-labetls for novel object categories. To simultaneously generate high-quality proposals and pseudo-labels, we propose CastDet, a CLIP-activated student-teacher open-vocabulary object Detection framework. Our end-to-end framework following the student-teacher self-learning mechanism employs the RemoteCLIP model as an extra omniscient teacher with rich knowledge. By doing so, our approach boosts not only novel object proposals but also classification. Furthermore, we devise a dynamic label queue strategy to maintain high-quality pseudo labels during batch training. We conduct extensive experiments on multiple existing aerial object detection datasets, which are set up for the OVD task. Experimental results demonstrate our CastDet achieving superior open-vocabulary detection performance, e.g., reaching 46.5% mAP on VisDroneZSD novel categories, which outperforms the state-of-the-art open-vocabulary detectors by 21.0% mAP. To our best knowledge, this is the first work to apply and develop the open-vocabulary object detection technique for aerial images. The code is available at https:\/\/github.com\/lizzy8587\/CastDet.\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-5492ad19 e-flex e-con-boxed e-con e-parent\" data-id=\"5492ad19\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-6af720c0 elementor-widget elementor-widget-spacer\" data-id=\"6af720c0\" data-element_type=\"widget\" data-widget_type=\"spacer.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor - v3.22.0 - 24-06-2024 *\/\n.elementor-column .elementor-spacer-inner{height:var(--spacer-size)}.e-con{--container-widget-width:100%}.e-con-inner>.elementor-widget-spacer,.e-con>.elementor-widget-spacer{width:var(--container-widget-width,var(--spacer-size));--align-self:var(--container-widget-align-self,initial);--flex-shrink:0}.e-con-inner>.elementor-widget-spacer>.elementor-widget-container,.e-con>.elementor-widget-spacer>.elementor-widget-container{height:100%;width:100%}.e-con-inner>.elementor-widget-spacer>.elementor-widget-container>.elementor-spacer,.e-con>.elementor-widget-spacer>.elementor-widget-container>.elementor-spacer{height:100%}.e-con-inner>.elementor-widget-spacer>.elementor-widget-container>.elementor-spacer>.elementor-spacer-inner,.e-con>.elementor-widget-spacer>.elementor-widget-container>.elementor-spacer>.elementor-spacer-inner{height:var(--container-widget-height,var(--spacer-size))}.e-con-inner>.elementor-widget-spacer.elementor-widget-empty,.e-con>.elementor-widget-spacer.elementor-widget-empty{position:relative;min-height:22px;min-width:22px}.e-con-inner>.elementor-widget-spacer.elementor-widget-empty .elementor-widget-empty-icon,.e-con>.elementor-widget-spacer.elementor-widget-empty .elementor-widget-empty-icon{position:absolute;top:0;bottom:0;left:0;right:0;margin:auto;padding:0;width:22px;height:22px}<\/style>\t\t<div class=\"elementor-spacer\">\n\t\t\t<div class=\"elementor-spacer-inner\"><\/div>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-48ddd0cd e-flex e-con-boxed e-con e-parent\" data-id=\"48ddd0cd\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-7910e744 elementor-widget elementor-widget-heading\" data-id=\"7910e744\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h2 class=\"elementor-heading-title elementor-size-default\">\u76f8\u5173\u4fe1\u606f\/Info<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-119116f8 e-flex e-con-boxed e-con e-parent\" data-id=\"119116f8\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t<div class=\"elementor-element elementor-element-24c303b3 e-con-full e-flex e-con e-child\" data-id=\"24c303b3\" data-element_type=\"container\">\n\t\t\t\t<div class=\"elementor-element elementor-element-928a530 elementor-widget elementor-widget-heading\" data-id=\"928a530\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h2 class=\"elementor-heading-title elementor-size-default\">\u4f5c\u8005\/Authors<br><\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-21be061b elementor-widget elementor-widget-heading\" data-id=\"21be061b\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h2 class=\"elementor-heading-title elementor-size-default\">\u94fe\u63a5\/Link<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-1a7b139c e-con-full e-flex e-con e-child\" data-id=\"1a7b139c\" data-element_type=\"container\">\n\t\t\t\t<div class=\"elementor-element elementor-element-3a226f46 elementor-widget elementor-widget-text-editor\" data-id=\"3a226f46\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>Yan Li, Weiwei Guo, Xue Yang, Ning Liao, Dunyun He, Jiaqi Zhou, Wenxian Yu<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<div class=\"elementor-element elementor-element-a4bfd54 elementor-widget elementor-widget-text-editor\" data-id=\"a4bfd54\" data-element_type=\"widget\" data-widget_type=\"text-editor.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t\t\t<p>https:\/\/link.springer.com\/chapter\/10.1007\/978-3-031-73016-0_25<\/p>\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-3e533acf e-flex e-con-boxed e-con e-parent\" data-id=\"3e533acf\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-6060fe2b elementor-widget elementor-widget-spacer\" data-id=\"6060fe2b\" data-element_type=\"widget\" data-widget_type=\"spacer.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-spacer\">\n\t\t\t<div class=\"elementor-spacer-inner\"><\/div>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-3a43c0e5 e-flex e-con-boxed e-con e-parent\" data-id=\"3a43c0e5\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-22b8f28b elementor-widget elementor-widget-spacer\" data-id=\"22b8f28b\" data-element_type=\"widget\" data-widget_type=\"spacer.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t\t\t<div class=\"elementor-spacer\">\n\t\t\t<div class=\"elementor-spacer-inner\"><\/div>\n\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-6dea13ae e-flex e-con-boxed e-con e-parent\" data-id=\"6dea13ae\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-796b2ec elementor-widget elementor-widget-heading\" data-id=\"796b2ec\" data-element_type=\"widget\" data-widget_type=\"heading.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<h2 class=\"elementor-heading-title elementor-size-default\">\u56fe\u7247\/Figures<\/h2>\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t<div class=\"elementor-element elementor-element-28a61b33 e-flex e-con-boxed e-con e-parent\" data-id=\"28a61b33\" data-element_type=\"container\">\n\t\t\t\t\t<div class=\"e-con-inner\">\n\t\t\t\t<div class=\"elementor-element elementor-element-78167274 elementor-widget elementor-widget-gallery\" data-id=\"78167274\" data-element_type=\"widget\" data-settings=\"{&quot;gallery_layout&quot;:&quot;masonry&quot;,&quot;columns&quot;:2,&quot;gap&quot;:{&quot;unit&quot;:&quot;px&quot;,&quot;size&quot;:20,&quot;sizes&quot;:[]},&quot;lazyload&quot;:&quot;yes&quot;,&quot;columns_tablet&quot;:2,&quot;columns_mobile&quot;:1,&quot;gap_tablet&quot;:{&quot;unit&quot;:&quot;px&quot;,&quot;size&quot;:10,&quot;sizes&quot;:[]},&quot;gap_mobile&quot;:{&quot;unit&quot;:&quot;px&quot;,&quot;size&quot;:10,&quot;sizes&quot;:[]},&quot;link_to&quot;:&quot;file&quot;,&quot;overlay_background&quot;:&quot;yes&quot;,&quot;content_hover_animation&quot;:&quot;fade-in&quot;}\" data-widget_type=\"gallery.default\">\n\t\t\t\t<div class=\"elementor-widget-container\">\n\t\t\t<style>\/*! elementor-pro - v3.22.0 - 24-06-2024 *\/\n.elementor-gallery__container{min-height:1px}.elementor-gallery-item{position:relative;overflow:hidden;display:block;text-decoration:none;border:solid var(--image-border-width) var(--image-border-color);border-radius:var(--image-border-radius)}.elementor-gallery-item__content,.elementor-gallery-item__overlay{height:100%;width:100%;position:absolute;top:0;left:0}.elementor-gallery-item__overlay{mix-blend-mode:var(--overlay-mix-blend-mode);transition-duration:var(--overlay-transition-duration);transition-property:mix-blend-mode,transform,opacity,background-color}.elementor-gallery-item__image.e-gallery-image{transition-duration:var(--image-transition-duration);transition-property:filter,transform}.elementor-gallery-item__content{display:flex;flex-direction:column;justify-content:var(--content-justify-content,center);align-items:center;text-align:var(--content-text-align);padding:var(--content-padding)}.elementor-gallery-item__content>div{transition-duration:var(--content-transition-duration)}.elementor-gallery-item__content.elementor-gallery--sequenced-animation>div:nth-child(2){transition-delay:calc(var(--content-transition-delay) \/ 3)}.elementor-gallery-item__content.elementor-gallery--sequenced-animation>div:nth-child(3){transition-delay:calc(var(--content-transition-delay) \/ 3 * 2)}.elementor-gallery-item__content.elementor-gallery--sequenced-animation>div:nth-child(4){transition-delay:calc(var(--content-transition-delay) \/ 3 * 3)}.elementor-gallery-item__description{color:var(--description-text-color,#fff);width:100%}.elementor-gallery-item__title{color:var(--title-text-color,#fff);font-weight:700;width:100%}.elementor-gallery__titles-container{display:flex;flex-wrap:wrap;justify-content:var(--titles-container-justify-content,center);margin-bottom:20px}.elementor-gallery__titles-container:not(.e--pointer-framed) .elementor-item:after,.elementor-gallery__titles-container:not(.e--pointer-framed) .elementor-item:before{background-color:var(--galleries-pointer-bg-color-hover)}.elementor-gallery__titles-container:not(.e--pointer-framed) .elementor-item.elementor-item-active:after,.elementor-gallery__titles-container:not(.e--pointer-framed) .elementor-item.elementor-item-active:before{background-color:var(--galleries-pointer-bg-color-active)}.elementor-gallery__titles-container.e--pointer-framed .elementor-item:before{border-color:var(--galleries-pointer-bg-color-hover);border-width:var(--galleries-pointer-border-width)}.elementor-gallery__titles-container.e--pointer-framed .elementor-item:after{border-color:var(--galleries-pointer-bg-color-hover)}.elementor-gallery__titles-container.e--pointer-framed .elementor-item.elementor-item-active:after,.elementor-gallery__titles-container.e--pointer-framed .elementor-item.elementor-item-active:before{border-color:var(--galleries-pointer-bg-color-active)}.elementor-gallery__titles-container.e--pointer-framed.e--animation-draw .elementor-item:before{border-left-width:var(--galleries-pointer-border-width);border-bottom-width:var(--galleries-pointer-border-width);border-right-width:0;border-top-width:0}.elementor-gallery__titles-container.e--pointer-framed.e--animation-draw .elementor-item:after{border-left-width:0;border-bottom-width:0;border-right-width:var(--galleries-pointer-border-width);border-top-width:var(--galleries-pointer-border-width)}.elementor-gallery__titles-container.e--pointer-framed.e--animation-corners .elementor-item:before{border-left-width:var(--galleries-pointer-border-width);border-bottom-width:0;border-right-width:0;border-top-width:var(--galleries-pointer-border-width)}.elementor-gallery__titles-container.e--pointer-framed.e--animation-corners .elementor-item:after{border-left-width:0;border-bottom-width:var(--galleries-pointer-border-width);border-right-width:var(--galleries-pointer-border-width);border-top-width:0}.elementor-gallery__titles-container .e--pointer-double-line .elementor-item:after,.elementor-gallery__titles-container .e--pointer-double-line .elementor-item:before,.elementor-gallery__titles-container .e--pointer-overline .elementor-item:before,.elementor-gallery__titles-container .e--pointer-underline .elementor-item:after{height:var(--galleries-pointer-border-width)}.elementor-gallery-title{--space-between:10px;cursor:pointer;color:#6d7882;font-weight:500;position:relative;padding:7px 14px;transition:all .3s}.elementor-gallery-title--active{color:#495157}.elementor-gallery-title:not(:last-child){margin-inline-end:var(--space-between)}.elementor-gallery-item__title+.elementor-gallery-item__description{margin-top:var(--description-margin-top)}.e-gallery-item.elementor-gallery-item{transition-property:all}.e-gallery-item.elementor-animated-content .elementor-animated-item--enter-from-bottom,.e-gallery-item.elementor-animated-content .elementor-animated-item--enter-from-left,.e-gallery-item.elementor-animated-content .elementor-animated-item--enter-from-right,.e-gallery-item.elementor-animated-content .elementor-animated-item--enter-from-top,.e-gallery-item:focus .elementor-gallery__item-overlay-bg,.e-gallery-item:focus .elementor-gallery__item-overlay-content,.e-gallery-item:focus .elementor-gallery__item-overlay-content__description,.e-gallery-item:focus .elementor-gallery__item-overlay-content__title,.e-gallery-item:hover .elementor-gallery__item-overlay-bg,.e-gallery-item:hover .elementor-gallery__item-overlay-content,.e-gallery-item:hover .elementor-gallery__item-overlay-content__description,.e-gallery-item:hover .elementor-gallery__item-overlay-content__title{opacity:1}a.elementor-item.elementor-gallery-title{color:var(--galleries-title-color-normal)}a.elementor-item.elementor-gallery-title.elementor-item-active,a.elementor-item.elementor-gallery-title.highlighted,a.elementor-item.elementor-gallery-title:focus,a.elementor-item.elementor-gallery-title:hover{color:var(--galleries-title-color-hover)}a.elementor-item.elementor-gallery-title.elementor-item-active{color:var(--gallery-title-color-active)}.e-con-inner>.elementor-widget-gallery,.e-con>.elementor-widget-gallery{width:var(--container-widget-width);--flex-grow:var(--container-widget-flex-grow)}<\/style>\t\t<div class=\"elementor-gallery__container\">\n\t\t\t\t\t\t\t<a class=\"e-gallery-item elementor-gallery-item elementor-animated-content\" href=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41.png\" data-elementor-open-lightbox=\"no\">\n\t\t\t\t\t<div class=\"e-gallery-image elementor-gallery-item__image\" data-thumbnail=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-41.png\" data-width=\"856\" data-height=\"648\" aria-label=\"\" role=\"img\" ><\/div>\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"elementor-gallery-item__overlay\"><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t<a class=\"e-gallery-item elementor-gallery-item elementor-animated-content\" href=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-58.png\" data-elementor-open-lightbox=\"no\">\n\t\t\t\t\t<div class=\"e-gallery-image elementor-gallery-item__image\" data-thumbnail=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-10-58.png\" data-width=\"1726\" data-height=\"575\" aria-label=\"\" role=\"img\" ><\/div>\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"elementor-gallery-item__overlay\"><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t<a class=\"e-gallery-item elementor-gallery-item elementor-animated-content\" href=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-20.png\" data-elementor-open-lightbox=\"no\">\n\t\t\t\t\t<div class=\"e-gallery-image elementor-gallery-item__image\" data-thumbnail=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-20.png\" data-width=\"1726\" data-height=\"860\" aria-label=\"\" role=\"img\" ><\/div>\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"elementor-gallery-item__overlay\"><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t<a class=\"e-gallery-item elementor-gallery-item elementor-animated-content\" href=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-45.png\" data-elementor-open-lightbox=\"no\">\n\t\t\t\t\t<div class=\"e-gallery-image elementor-gallery-item__image\" data-thumbnail=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-45.png\" data-width=\"1748\" data-height=\"431\" aria-label=\"\" role=\"img\" ><\/div>\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"elementor-gallery-item__overlay\"><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t\t\t<a class=\"e-gallery-item elementor-gallery-item elementor-animated-content\" href=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-58.png\" data-elementor-open-lightbox=\"no\">\n\t\t\t\t\t<div class=\"e-gallery-image elementor-gallery-item__image\" data-thumbnail=\"https:\/\/tongjicdi.com\/wp-content\/uploads\/2025\/05\/Snipaste_2025-05-29_20-11-58.png\" data-width=\"1726\" data-height=\"594\" aria-label=\"\" role=\"img\" ><\/div>\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"elementor-gallery-item__overlay\"><\/div>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/a>\n\t\t\t\t\t<\/div>\n\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"<p>An increasingly massive number of remote-sensing images spurs the development of extensible object detectors that can detect objects beyond training categories without costly collecting new labeled data. In this paper, we aim to develop open-vocabulary object detection (OVD) technique in aerial images that scales up object vocabulary size beyond training data. The performance of OVD greatly relies on the quality of class-agnostic region proposals and pseudo-labetls for novel object categories. To simultaneously generate high-quality proposals and pseudo-labels, we propose CastDet, a CLIP-activated student-teacher open-vocabulary object Detection framework. Our end-to-end framework following the student-teacher self-learning mechanism employs the RemoteCLIP model as an extra omniscient teacher with rich knowledge. By doing so, our approach boosts not only novel object proposals but also classification. Furthermore, we devise a dynamic label queue strategy to maintain high-quality pseudo labels during batch training. We conduct extensive experiments on multiple existing aerial object detection datasets, which are set up for the OVD task. Experimental results demonstrate our CastDet achieving superior open-vocabulary detection performance, e.g., reaching 46.5% mAP on VisDroneZSD novel categories, which outperforms the state-of-the-art open-vocabulary detectors by 21.0% mAP. To our best knowledge, this is the first work to apply and develop the open-vocabulary object detection technique for aerial images. The code is available at https:\/\/github.com\/lizzy8587\/CastDet.<\/p>\n","protected":false},"author":1,"featured_media":1675,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[29],"tags":[],"class_list":["post-1674","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-cv"],"_links":{"self":[{"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/posts\/1674","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/comments?post=1674"}],"version-history":[{"count":4,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/posts\/1674\/revisions"}],"predecessor-version":[{"id":1684,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/posts\/1674\/revisions\/1684"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/media\/1675"}],"wp:attachment":[{"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/media?parent=1674"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/categories?post=1674"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/tongjicdi.com\/index.php\/wp-json\/wp\/v2\/tags?post=1674"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}