{"id":13111,"date":"2026-10-09T22:46:16","date_gmt":"2026-10-09T14:46:16","guid":{"rendered":"https:\/\/infernews.com\/blog\/unified-image-evaluation-with-spatially-grounded-explanations\/"},"modified":"2026-10-09T22:50:39","modified_gmt":"2026-10-09T14:50:39","slug":"viescore2","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/viescore2\/","title":{"rendered":"VIEScore2 \u8b93\u5f71\u50cf\u8a55\u5206\u4e0d\u518d\u53ea\u5269\u4e00\u500b\u5206\u6578"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/10\/VIEScore2.jpg\" alt=\"teaser\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u8a55\u4f30 AI \u751f\u6210\u6216\u7de8\u8f2f\u5f71\u50cf\u6642\uff0c\u55ae\u4e00\u5206\u6578\u5f80\u5f80\u53ea\u80fd\u544a\u8a34\u4f60\u597d\u4e0d\u597d\uff0c\u537b\u8aaa\u4e0d\u51fa\u554f\u984c\u5728\u54ea\u88cf\u3002VIEScore2 \u5c6c\u65bc\u5f71\u50cf\u8a55\u4f30\u6a21\u578b\uff0c\u8655\u7406\u751f\u6210\u5716\u3001\u7de8\u8f2f\u5716\u3001\u63d0\u793a\u8a5e\u53ca\u53ef\u9078\u7684\u689d\u4ef6\u5f71\u50cf\uff0c\u4e26\u5728\u4e00\u6b21\u6a21\u578b\u904b\u7b97\u4e2d\u8f38\u51fa\u8cea\u7d20\u3001\u8a9e\u610f\u4e00\u81f4\u6027\u548c\u7f3a\u9677\u4f4d\u7f6e\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Qwen3-VL-8B \u7d93\u5fae\u8abf\u5f8c\uff0c\u4ee5\u81ea\u56de\u6b78\u65b9\u5f0f\u540c\u6642\u9810\u6e2c perceptual-quality\uff08PQ\uff09\u3001semantic-consistency\uff08SC\uff09\u53ca 16\u00d716 defect grid\uff1b<code>artifact<\/code> \u548c <code>misalign<\/code> \u5169\u500b\u901a\u9053\u53ef\u5206\u958b\u6a19\u793a\u8996\u89ba\u7455\u75b5\u8207\u8a9e\u610f\u932f\u4f4d\u3002\u53c3\u6578\u514d\u8a13\u7df4\u7684 parser \u518d\u628a\u7db2\u683c\u7d50\u679c\u7a69\u5b9a\u8f49\u6210\u6587\u5b57\u89e3\u91cb\uff0c\u907f\u514d\u8aaa\u660e\u5167\u5bb9\u8207\u5b9a\u4f4d\u7d50\u679c\u4e92\u76f8\u77db\u76fe\u3002<\/p>\n\n\n<div\n    \tclass=\"align wp-block-vpb-video\"    id='vpbpVideoPlayer-1'\n    data-attributes='{&quot;source&quot;:&quot;https:\\\/\\\/tiger-ai-lab.github.io\\\/VIEScore2\\\/static\\\/videos\\\/demo.mp4&quot;,&quot;repeat&quot;:true,&quot;autoplay&quot;:true,&quot;muted&quot;:true,&quot;align&quot;:&quot;&quot;,&quot;poster&quot;:&quot;&quot;,&quot;controls&quot;:{&quot;play-large&quot;:true,&quot;restart&quot;:false,&quot;rewind&quot;:true,&quot;play&quot;:true,&quot;fast-forward&quot;:true,&quot;progress&quot;:true,&quot;current-time&quot;:true,&quot;duration&quot;:false,&quot;mute&quot;:true,&quot;volume&quot;:true,&quot;pip&quot;:false,&quot;airplay&quot;:false,&quot;captions&quot;:false,&quot;settings&quot;:true,&quot;download&quot;:false,&quot;fullscreen&quot;:true},&quot;captions&quot;:[],&quot;ratio&quot;:&quot;16:9&quot;,&quot;accentColor&quot;:&quot;&quot;,&quot;lazyLoad&quot;:false,&quot;width&quot;:&quot;100%&quot;,&quot;radius&quot;:&quot;0px&quot;,&quot;resetOnEnd&quot;:false,&quot;autoHideControl&quot;:true,&quot;isSetup&quot;:false}'\n><\/div>\n\n\n<p class=\"wp-block-paragraph\">\u76f8\u6bd4\u53ea\u8f38\u51fa\u7e3d\u5206\u7684\u8a55\u4f30\u5668\uff0cVIEScore2 \u63d0\u4f9b\u8f03\u5177\u64cd\u4f5c\u6027\u7684\u7a7a\u9593\u7dda\u7d22\uff1b\u76f8\u6bd4\u901a\u7528 VLM\uff0c\u6a21\u578b\u5247\u5c08\u6ce8\u65bc\u5f71\u50cf\u8a55\u5206\u548c\u7f3a\u9677\u5b9a\u4f4d\uff0c\u4ee3\u50f9\u662f\u7db2\u683c\u4ee5 16\u00d716 \u55ae\u5143\u8868\u9054\uff0c\u7d30\u7bc0\u7cbe\u5ea6\u6709\u9650\u3002\u8a13\u7df4\u8cc7\u6599\u628a ImagenWorld\u3001RichHF-18K\u3001PAL4VST\u3001COCO \u7684\u906e\u7f69\u6216\u71b1\u5716\uff0c\u4ee5\u53ca EvalMuse-40K \u7684\u5206\u6578\u8cc7\u6599\u7d71\u4e00\u6210\u540c\u4e00\u683c\u5f0f\uff0c\u518d\u4ee5 SFT \u548c GRPO \u914d\u5408 cell-level Dice\u3001\u5206\u6578\u6e96\u78ba\u5ea6\u53ca\u683c\u5f0f\u6709\u6548\u6027\u734e\u52f5\u8a13\u7df4\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">GitHub \u5132\u5b58\u5eab\u63d0\u4f9b\u8a55\u4f30\u5354\u8b70\u3001\u8cc7\u6599 manifests\u3001\u5916\u90e8\u6e2c\u8a66\u96c6\u8f49\u63db\u5668\uff0c\u4ee5\u53ca <code>run_eval.py<\/code>\uff1b\u53ef\u4f7f\u7528\u672c\u5730 checkpoint\uff0c\u4ea6\u652f\u63f4 OpenAI\u3001Gemini \u548c Anthropic API backend\u3002\u4e3b\u8981\u5957\u4ef6\u5305\u62ec AbHuman\u3001AGIQA-3K\u3001EvalMuse\u3001PAL4VST \u7b49\u8a55\u4f30\u96c6\uff0c\u9069\u5408\u7814\u7a76\u5f71\u50cf\u751f\u6210\u6a21\u578b\u3001\u5efa\u7acb\u81ea\u52d5\u8cea\u91cf\u6aa2\u67e5\u6d41\u7a0b\uff0c\u6216\u5206\u6790\u7de8\u8f2f\u7d50\u679c\u7684\u932f\u8aa4\u4f4d\u7f6e\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u8a55\u5206\u66f4\u5b8c\u6574<\/strong>\uff1a\u540c\u6642\u8f38\u51fa PQ\u3001SC\u3001\u7f3a\u9677\u7db2\u683c\u53ca\u6587\u5b57\u89e3\u91cb\u3002<\/li>\n\n\n\n<li><strong>\u7d50\u679c\u8f03\u63a5\u8fd1\u4eba\u8a55<\/strong>\uff1a\u4e3b\u8981\u6e2c\u8a66\u96c6 overall-score SRCC \u70ba 0.601\uff0c\u6700\u4f73 VLM baseline \u70ba 0.491\u3002<\/li>\n\n\n\n<li><strong>\u5b9a\u4f4d\u80fd\u529b\u5177\u7af6\u722d\u529b<\/strong>\uff1a6 \u500b\u5b9a\u4f4d\u57fa\u6e96\u4e2d\u6709 3 \u500b\u53d6\u5f97\u6700\u9ad8 grid IoU\uff0c5 \u500b\u9032\u5165\u524d\u4e09\u3002<\/li>\n\n\n\n<li><strong>\u9650\u5236\u6e05\u6670<\/strong>\uff1a\u7db2\u683c\u5b9a\u4f4d\u4ecd\u5c6c\u5340\u57df\u7d1a\u63d0\u793a\uff0c\u4e26\u9700\u8981\u6309\u5132\u5b58\u5eab\u5354\u8b70\u6574\u7406\u8cc7\u6599\u8207\u8f38\u5165\u683c\u5f0f\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/tiger-ai-lab.github.io\/VIEScore2\/\" rel=\"noopener noreferrer\"><strong>\u9805\u76ee\u4e3b\u9801<\/strong><\/a> \u00b7 <a href=\"https:\/\/github.com\/TIGER-AI-Lab\/VIEScore2\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a> \u00b7 <a href=\"https:\/\/huggingface.co\/Allenda\/VIEScore2\" rel=\"noopener noreferrer\"><strong>\u6a21\u578b<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>VIEScore2 \u540c\u6642\u8a55\u4f30\u5f71\u50cf\u8cea\u7d20\u3001\u8a9e\u610f\u4e00\u81f4\u6027\u8207\u7f3a\u9677\u4f4d\u7f6e\uff0c\u5c07\u96e3\u4ee5\u89e3\u8b80\u7684\u5206\u6578\u8f49\u6210\u53ef\u8ffd\u8e64\u7684\u4fee\u6539\u65b9\u5411\u3002<\/p>\n","protected":false},"author":8,"featured_media":13110,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"wpai_generated_summary":"","wpai_meta_description":"","footnotes":""},"categories":[133,76,149,127,185,179,153,140,137,30,161,150,191],"tags":[],"class_list":["post-13111","post","type-post","status-publish","format-standard","hentry","category-133","category-76","category-149","category-127","category-qwen","category-nvidia","category-openai","category-gemini","category-api","category-image","category-161","category-150","category-anthropic"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/13111","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=13111"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/13111\/revisions"}],"predecessor-version":[{"id":13113,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/13111\/revisions\/13113"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/13110"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=13111"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=13111"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=13111"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}