{"id":9438,"date":"2026-06-21T04:55:09","date_gmt":"2026-06-20T20:55:09","guid":{"rendered":"https:\/\/infernews.com\/blog\/adavomp-3d\/"},"modified":"2026-06-21T04:56:43","modified_gmt":"2026-06-20T20:56:43","slug":"adavomp-3d","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/adavomp-3d\/","title":{"rendered":"AdaVoMP \u8b93 3D \u5834\u666f\u6709\u53ef\u4e92\u52d5\u7269\u7406\u6750\u8cea"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/06\/Capture3_medium.jpg\" alt=\"Og image\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">\u4e0d\u5c11 3D \u8cc7\u7522\u53ea\u6709\u5916\u5f62\uff0c\u6b20\u7f3a Young&#8217;s modulus\uff08E\uff09\u3001Poisson&#8217;s ratio\uff08\u03bd\uff09\u540c density\uff08\u03c1\uff09\u7b49\u7269\u7406\u8cc7\u6599\uff0c\u7d50\u679c\u505a\u6a21\u64ec\u6642\u53ea\u4fc2\u300c\u7747\u843d\u4f3c\u300d\uff0c\u4f46\u53d7\u529b\u3001\u8b8a\u5f62\u540c\u78b0\u649e\u672a\u5fc5\u53ef\u4fe1\u3002AdaVoMP \u5c31\u4fc2\u91dd\u5c0d\u5462\u500b\u7f3a\u53e3\uff0c\u70ba\u8f38\u5165 3D \u7269\u4ef6\u9810\u6e2c\u9ad8\u5bc6\u5ea6\u3001\u7a7a\u9593\u53ef\u8b8a\u7684\u7269\u7406\u5c6c\u6027\u5834\uff0c\u4ee4\u6578\u78bc\u4e16\u754c\u66f4\u63a5\u8fd1\u53ef\u4e92\u52d5\u3001\u53ef\u6a21\u64ec\u7684\u72c0\u614b\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u540c\u524d\u4e00\u4ee3 VoMP v1 \u6bd4\uff0cAdaVoMP \u7684\u91cd\u9ede\u5514\u6b62\u4fc2\u66f4\u6e05\u6670\uff0c\u800c\u4fc2\u6539\u7528 sparse and adaptive voxel structure\uff08SAV\uff09\u540c sparse transformer encoder-decoder\uff0c\u5c07\u56fa\u5b9a\u9ad4\u7d20\u8868\u793a\u63db\u6210\u53ef\u81ea\u9069\u61c9\u7d50\u69cb\u3002\u9801\u9762\u6307\u51fa\uff0c\u5b83\u53ef\u751f\u6210\u9ad8\u51fa 16^3 \u500d\u89e3\u6790\u5ea6\u7684\u5c6c\u6027\u5834\uff0c\u4ea6\u652f\u63f4 test-time \u8abf\u6574\u89e3\u6790\u5ea6\uff0c\u517c\u9867\u6e96\u78ba\u5ea6\u3001\u8a18\u61b6\u9ad4\u6548\u7387\u540c\u7d30\u7bc0\u4fdd\u7559\u3002<\/p>\n\n\n<figure class=\"wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"lyte-wrapper\" title=\"AdaVoMP short video [ICML 26]\" style=\"width:853px;max-width:100%;margin:5px auto;\"><div class=\"lyMe\" id=\"WYL_a5sXqAWlbEs\" itemprop=\"video\" itemscope itemtype=\"https:\/\/schema.org\/VideoObject\"><div><meta itemprop=\"thumbnailUrl\" content=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fa5sXqAWlbEs%2Fhqdefault.jpg\" \/><meta itemprop=\"embedURL\" content=\"https:\/\/www.youtube.com\/embed\/a5sXqAWlbEs\" \/><meta itemprop=\"duration\" content=\"PT49S\" \/><meta itemprop=\"uploadDate\" content=\"2026-06-19T23:46:54Z\" \/><\/div><div id=\"lyte_a5sXqAWlbEs\" data-src=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fa5sXqAWlbEs%2Fhqdefault.jpg\" class=\"pL\"><div class=\"tC\"><div class=\"tT\" itemprop=\"name\">AdaVoMP short video [ICML 26]<\/div><\/div><div class=\"play\"><\/div><div class=\"ctrl\"><div class=\"Lctrl\"><\/div><div class=\"Rctrl\"><\/div><\/div><\/div><noscript><a href=\"https:\/\/youtu.be\/a5sXqAWlbEs\" rel=\"nofollow noopener\"><img loading=\"lazy\" decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/plugins\/wp-youtube-lyte\/lyteCache.php?origThumbUrl=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fa5sXqAWlbEs%2F0.jpg\" alt=\"AdaVoMP short video [ICML 26]\" width=\"853\" height=\"460\" \/><br \/>Watch this video on YouTube<\/a><\/noscript><meta itemprop=\"description\" content=\"Accurate mechanical properties (or materials) Young&#039;s modulus (&#36;E&#36;), Poisson&#039;s ratio (&#36;\\nu&#36;) and density (&#36;\\rho&#36;) are essential for reliable physics simulation of digital worlds, but most 3D assets lack this information. We propose AdaVoMP, a method for predicting accurate dense spatially-varying (&#36;E&#36;, &#36;\\nu&#36;, &#36;\\rho&#36;) for input 3D objects across representations, improving the resolution, accuracy, and memory efficiency over the state-of-the-art. The foundation of our technique is a sparse and adaptive voxel structure SAV that efficiently represents both the input 3D shape and the material field output. We replace the fixed-voxel model of the most accurate prior method, VoMP, with a novel sparse transformer encoder-decoder model that learns to generate a unique SAV autoregressively for every input shape to represent its materials, achieving a resolution &#36;16^3\\times&#36; higher than prior art. Experiments show that AdaVoMP estimates more accurate volumetric properties, even with lesser test-time compute than all prior art. This allows us to convert high-resolution complex 3D objects into simulation-ready assets, resulting in realistic deformable simulations.\"><\/div><\/div><div class=\"lL\" style=\"max-width:100%;width:853px;margin:5px auto;\"><\/div><figcaption><\/figcaption><\/figure>\n\n\n<p class=\"wp-block-paragraph\">AdaVoMP \u5c55\u793a\u5497\u5e7e\u7a2e\u76f4\u63a5\u53ef\u7406\u89e3\u7684\u7528\u9014\uff0c\u4f8b\u5982\u5c07 Gaussian Splat + mesh \u5834\u666f\u8f49\u6210\u53ef\u4e92\u52d5\u4e16\u754c\u3001\u70ba\u7d14 mesh \u6216 gaussian splats \u5834\u666f\u505a\u8f03\u771f\u5be6\u7684\u7269\u7406\u6a21\u64ec\uff0c\u4ee5\u53ca\u914d\u5408 RoboLab \u8207 Isaac Sim \u5efa\u7acb\u6a5f\u68b0\u4eba\u6e2c\u8a66\u74b0\u5883\u3002\u5c0d\u505a robotics\u3001simulation\u3001\u6578\u78bc\u5b7f\u751f\uff0c\u6216\u8005\u60f3\u5c07 NeRF\u3001Gaussian Splat \u8cc7\u7522\u8b8a\u6210\u53ef\u64cd\u4f5c\u5834\u666f\u7684\u4eba\uff0c\u5462\u985e\u6d41\u7a0b\u6703\u7279\u5225\u6709\u53c3\u8003\u50f9\u503c\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u88dc\u56de 3D \u8cc7\u7522\u5e38\u898b\u7f3a\u5931\u7684\u7269\u7406\u6750\u8cea\u8cc7\u8a0a<\/li>\n\n\n\n<li>\u6bd4 VoMP v1 \u63d0\u4f9b\u66f4\u9ad8\u89e3\u6790\u5ea6\u8207\u53ef\u7e2e\u653e\u8f38\u51fa<\/li>\n\n\n\n<li>\u540c\u6642\u652f\u63f4 mesh\u3001Gaussian Splat \u7b49\u4e0d\u540c\u8868\u793a<\/li>\n\n\n\n<li>\u53ef\u7528\u65bc Isaac Sim \u6a5f\u68b0\u4eba\u57fa\u6e96\u6e2c\u8a66\u8207\u4e92\u52d5\u5834\u666f\u5efa\u7acb<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u5f9e\u9801\u9762\u8cc7\u6599\u770b\uff0cAdaVoMP \u4ea6\u6709\u540c NeRF2Physics\u3001PUGS\u3001Phys4DGen\u3001Pixie\u3001VoMP \u53ca Ground Truth \u4f5c\u8996\u89ba\u6bd4\u8f03\uff0c\u91cd\u9ede\u653e\u5728 physics material fields \u7684\u54c1\u8cea\u3002\u82e5\u8981\u81ea\u884c\u6e2c\u8a66\uff0c\u67e5\u770b Code \u540c Model\/Data\uff0c\u518d\u7559\u610f\u5b83\u5728\u81ea\u5df1\u5834\u666f\u8868\u793a\u3001\u6a21\u64ec\u5668\u540c\u8cc7\u7522\u683c\u5f0f\u4e0a\u7684\u63a5\u5165\u6210\u672c\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u9805\u76ee\uff1a<\/strong> <a href=\"https:\/\/research.nvidia.com\/labs\/sil\/projects\/adavomp\/\" rel=\"noopener noreferrer\">https:\/\/research.nvidia.com\/labs\/sil\/projects\/adavomp\/<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>AdaVoMP \u6703\u70ba 3D \u7269\u4ef6\u88dc\u4e0a\u7269\u7406\u6750\u8cea\u8cc7\u6599\uff0c\u4ee4\u6a21\u64ec\u66f4\u771f\u5be6\uff0c\u4ea6\u66f4\u9ad8\u89e3\u50cf\u8207\u7701\u8a18\u61b6\u9ad4\u3002<\/p>\n","protected":false},"author":8,"featured_media":9437,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,179,132,76,184],"tags":[],"class_list":["post-9438","post","type-post","status-publish","format-standard","hentry","category-133","category-nvidia","category-3d","category-76","category-robotic"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9438","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=9438"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9438\/revisions"}],"predecessor-version":[{"id":9440,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/9438\/revisions\/9440"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/9437"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=9438"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=9438"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=9438"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}