{"id":9838,"date":"2026-07-03T21:49:13","date_gmt":"2026-07-03T13:49:13","guid":{"rendered":"https:\/\/infernews.com\/blog\/anygroundbench\/"},"modified":"2026-07-03T21:49:13","modified_gmt":"2026-07-03T13:49:13","slug":"anygroundbench","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/anygroundbench\/","title":{"rendered":"AnyGroundBench \u9ede\u51fa\u5f71\u7247\u5b9a\u4f4d\u6a21\u578b\u76f2\u9ede"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/07\/pasted-a98923bdaae6.jpg\" alt=\"Repository image for rinost081\/AnyGroundBench\"><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">AnyGroundBench \u662f\u4e00\u500b\u5f71\u7247 grounding benchmark\uff0c\u4e5f\u662f\u9762\u5411\u5c08\u696d\u9818\u57df\u7684\u8cc7\u6599\u96c6\u8207\u8a55\u6e2c\u57fa\u6e96\u3002\u5b83\u4e3b\u8981\u7528\u4f86\u6e2c\u8a66 Vision-Language Models\uff08VLMs\uff09\u5728 animal\u3001industry\u3001sports\u3001surgery\u3001public security \u5e7e\u985e\u5834\u666f\u4e2d\uff0c\u80fd\u5426\u628a\u6587\u5b57\u63cf\u8ff0\u6e96\u78ba\u5c0d\u61c9\u5230\u5f71\u7247\u4e2d\u7684\u6642\u9593\u3001\u4f4d\u7f6e\uff0c\u4ee5\u53ca\u6642\u7a7a\u540c\u6642\u767c\u751f\u7684\u4e8b\u4ef6\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u73fe\u6709\u505a\u6cd5\u591a\u6578\u505c\u7559\u5728 general\u3001daily-life benchmark \u7684 zero-shot \u6e2c\u8a66\uff0c\u91cd\u9ede\u662f\u770b\u6a21\u578b\u6709\u6c92\u6709\u901a\u7528\u7406\u89e3\u529b\uff1b\u4f5c\u8005\u8a8d\u70ba\u9019\u7a2e\u7bc4\u5f0f\u7121\u6cd5\u53cd\u6620\u5c08\u9580\u5834\u666f\uff0c\u56e0\u70ba\u7a00\u6709\u8996\u89ba\u6982\u5ff5\u3001\u8907\u96dc\u52d5\u4f5c\u95dc\u4fc2\u8207\u9818\u57df\u8853\u8a9e\uff0c\u901a\u5e38\u4e0d\u6703\u5728\u901a\u7528\u8cc7\u6599\u88e1\u88ab\u5145\u5206\u5b78\u5230\u3002AnyGroundBench \u56e0\u800c\u628a\u8a55\u6e2c\u91cd\u5fc3\u8f49\u53bb domain adaptation\uff0c\u4e26\u52a0\u5165 dedicated training subsets\uff0c\u4ee4\u6e2c\u8a66\u4e0d\u518d\u53ea\u554f\u6a21\u578b\u300c\u6709\u6c92\u6709\u898b\u904e\u300d\uff0c\u800c\u662f\u9032\u4e00\u6b65\u91cf\u5ea6\u5b83\u300c\u80fd\u4e0d\u80fd\u9069\u61c9\u65b0\u9818\u57df\u300d\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u500b\u9805\u76ee\u7684\u5dee\u7570\uff0c\u5728\u65bc\u5b83\u628a temporal\u3001spatial\u3001spatio-temporal annotations \u7528\u7d71\u4e00\u65b9\u5f0f\u6574\u7406\uff0c\u4e26\u6df7\u5408 newly captured videos \u8207 existing datasets\u3002\u8cc7\u6599\u4f86\u6e90\u6db5\u84cb mouse\u3001american_football\u3001Animal-Kingdom\u3001MECCANO\u3001EgoSurgery \u7b49\uff0c\u8986\u84cb\u9762\u6bd4\u55ae\u4e00\u9818\u57df benchmark \u5ee3\uff0c\u4ea6\u66f4\u63a5\u8fd1\u7814\u7a76\u5718\u968a\u3001\u7522\u696d\u5206\u6790\u3001\u91ab\u7642\u5f71\u50cf\u7814\u7a76\u8207\u5b89\u5168\u76e3\u6e2c\u5834\u666f\u6703\u9047\u5230\u7684\u8cc7\u6599\u5206\u4f48\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9805\u76ee\u63d0\u4f9b Hugging Face dataset\u3001project page\uff1a\u9019\u4e0d\u662f\u5373\u63d2\u5373\u7528\u61c9\u7528\u7a0b\u5f0f\uff0c\u800c\u662f\u4f9b\u7814\u7a76\u8207\u6a21\u578b\u6bd4\u8f03\u7684 benchmark\u3002\u90e8\u7f72\u91cd\u9ede\u4e0d\u662f\u4ecb\u9762\u5b89\u88dd\uff0c\u800c\u662f\u6309 domain \u8b80\u53d6\u6574\u7406\u5f8c\u7684\u8cc7\u6599\uff0c\u7136\u5f8c\u4ee5 STVG\u3001TVG\u3001SVG \u4e09\u985e\u4efb\u52d9\u8dd1\u63a8\u7406\u8207\u8a55\u5206\uff1b\u6307\u6a19\u5206\u5225\u7528 vIoU@0.3\u3001tIoU@0.3\u3001sIoU@0.3\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u985e\u578b\u5c6c\u65bc <strong>benchmark \/ \u8cc7\u6599\u96c6<\/strong>\uff0c\u76ee\u7684\u662f\u6e2c\u91cf VLMs \u7684 specialized-domain video grounding \u80fd\u529b<\/li>\n<li>\u820a\u7bc4\u5f0f\u4ee5 zero-shot general benchmark \u70ba\u4e3b\uff0c\u65b0\u8a2d\u8a08\u6539\u70ba\u6aa2\u67e5 domain adaptation \u8207 In-Context Learning\uff08ICL\uff09\u662f\u5426\u771f\u7684\u6709\u6548<\/li>\n<li>\u8a55\u6e2c\u6db5\u84cb temporal\u3001spatial\u3001spatio-temporal \u4e09\u5c64\uff0c\u8f03\u5bb9\u6613\u770b\u51fa\u6a21\u578b\u7a76\u7adf\u662f\u770b\u932f\u6642\u9593\u3001\u627e\u932f\u4f4d\u7f6e\uff0c\u9084\u662f\u5169\u908a\u90fd\u5931\u6e96<\/li>\n<li>\u5df2\u8a55\u6e2c 15 \u500b state-of-the-art VLMs\uff0c\u7d50\u679c\u6307\u51fa\u73fe\u6709\u6a21\u578b\u5728 specialized domains \u7684 zero-shot \u8207 ICL \u8868\u73fe\u90fd\u4e0d\u7a69\u5b9a<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u5efa\u8b70\u6a21\u578b\u5305\u62ec GPT-4o\u3001GPT-5.1\u3001Gemini-2.5-Flash \u7b49 proprietary VLMs\uff1b\u73fe\u6709\u7d50\u679c\u986f\u793a\uff0c\u52a0\u5165 2-shot ICL \u96d6\u7136\u5728\u90e8\u5206 domain \u6709\u6539\u5584\uff0c\u4f46\u6574\u9ad4\u4ecd\u672a\u89e3\u6c7a specialized-domain spatio-temporal reasoning \u7684\u7f3a\u53e3\u3002\u5c0d\u7814\u7a76 VLM evaluation\u3001video grounding\u3001\u8996\u89ba\u6a21\u578b\u9077\u79fb\u80fd\u529b\u7684\u5718\u968a\u4f86\u8aaa\uff0c\u9019\u500b\u9805\u76ee\u6700\u6709\u50f9\u503c\u7684\u5730\u65b9\uff0c\u662f\u5b83\u628a\u300c\u901a\u7528\u6e2c\u8a66\u770b\u4f3c\u53ef\u7528\u300d\u8207\u300c\u5c08\u696d\u5834\u666f\u4ecd\u7136\u5931\u624b\u300d\u4e4b\u9593\u7684\u5dee\u8ddd\u91cf\u5316\u51fa\u4f86\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/rinost081.github.io\/AnyGroundBench-page\/\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/rinost081\/AnyGroundBench\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>GitHub<\/strong><\/a> \u00b7 <a href=\"https:\/\/arxiv.org\/pdf\/2607.02269\" rel=\"noopener noreferrer\" target=\"_blank\"><strong>Paper<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>AnyGroundBench\u4e0d\u662f\u65b0\u6a21\u578b\uff0c\u800c\u662f\u5c08\u9580\u6e2c\u8a66\u5f71\u7247\u7406\u89e3\u80fd\u529b\u7684 benchmark\u3002\u5b83\u628a\u5c08\u696d\u5834\u666f\u653e\u5165\u540c\u4e00\u5957\u898f\u5247\uff0c\u76f4\u63a5\u63ed\u793a\u73fe\u6709 VLMs \u7684\u9069\u61c9\u843d\u5dee\u3002<\/p>\n","protected":false},"author":8,"featured_media":9837,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,185,140,179,153,31,119,170,127,149,199],"tags":[],"class_list":["post-9838","post","type-post","status-publish","format-standard","hentry","category-133","category-qwen","category-gemini","category-nvidia","category-openai","category-video","category-119","category-170","category-127","category-149","category-dataset-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9838","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=9838"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9838\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/9837"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=9838"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=9838"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=9838"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}