[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-omnishowcase-360-subject-video":3,"topics-all":38,"news-related-b860aced-d186-4d56-b751-53d5ca1ed658":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"b860aced-d186-4d56-b751-53d5ca1ed658","OmniShowcase:4张参考图,主体转360度不变形","KAIST AI 团队发布 OmniShowcase:给 4 张参考图,视频模型即可 360 度环绕拍摄同一主体,身份与 3D 形状一致;靠几何驱动数据管线与表面法线转台条件,几何误差较最强基线砍半,数据集与代码暂未放出。","一段 360 度环绕商品的视频,如今只要 4 张参考图。10 月 26 日,KAIST AI 联合仁荷大学、FLIPTION 的团队把 OmniShowcase 挂上 arXiv(编号 2610.35684),三天后冲上 Hugging Face Daily Papers 10 月 29 日榜单第 2 名,拿到 42 个点赞。它瞄准的问题非常具体:主体到视频(S2V)生成里,镜头一旦转到参考图没拍到的背面,模型就开始\"编\"。\n\n## 单张参考图撑不起 360 度\n\nsubject-to-video 的常规做法是拿一张参考图,让视频模型在相机移动时自己推断主体看不见的部分。近来的多视角方法把参考扩到几张图,信息是多了,但没有显式告诉模型这几张图背后连接的 3D 结构是什么;更隐蔽的问题在训练数据本身——论文指出,现有数据不保证视角变化后仍是同一个 3D 主体,模型等于在学一门没有标准答案的课。\n\n## 两件武器:几何造数据,法线当转台\n\n第一件武器是全自动数据合成管线:从单个精选 3D 资产出发,用 Blender 同时渲染目标轨道视频和 4 张参考视角,再把目标主体合成进文本驱动的场景。关键细节是只对背景做 inpaint,主体背面永远来自真实渲染,不给模型\"幻觉\"的机会。由此得到 Showcase-4K 训练集:覆盖 12 种相机运动、88 种轨迹配置、约 4000 条视频,每条都带从重建导出的表面法线条件;另有留出的 ShowcaseBench 评测集,直接附上每条视频的源 mesh 与相机参数——这种\"自带标准答案\"的评测集在视频生成领域相当稀缺。\n\n第二件武器是几何条件:把 4 张参考图在 3D 里重建,渲染成一圈\"表面法线转台\",与参考图一起送进视频模型的参考通路。底座是 Wan2.1 图生视频扩散 Transformer,只做轻量 LoRA 微调;法线转台与生成帧之间不需要逐帧对齐,也不需要逐主体优化。转台本身是固定的偏航扫描,但生成相机并不被它绑死——轨迹由提示词决定,从 360 度环绕、螺旋下降到推拉摇移都能走。\n\n## 数字:几何误差砍半,美学不冒进\n\n在其自建 ShowcaseBench 上,OmniShowcase 身份分数 DINO 0.799、CLIP 0.925,几何指标 GT-CD 0.0824,在 Phantom-MV、MAGREF-MV、MV-S2V、3DreamBooth 四个多视角基线中,身份与几何相关指标全部居首;相对其中最强者,几何误差在留出集上约低 52%、在真实物体数据集 NAVI 上约低 46%——\"砍半\"的说法来自作者在论文页的自述,与项目页量化表可以互相验算。值得注意的是美学得分 0.574,低于 MAGREF-MV 的 0.614:几何一致性是硬赢,画面好看程度只算持平。消融同样做得到位:去掉法线、换成带纹理渲染、或抽稀转台,MEt3R 从 0.190 退化到 0.26 上下,证明法线条件确实是收益来源。\n\n## 冷水:论文先行,四件套一件没放\n\nGitHub 仓库 10 月 29 日才创建,目前只有一个 initial commit、8 个 star。README 的发布清单上,Showcase-4K、ShowcaseBench、推理代码、训练代码四项全部未勾选;Hugging Face 页面\"引用本论文的模型\"数量为 0。也就是说,论文与效果演示先行,可复现的全套工件都还在路上——急着拿数据集做 3D 一致性训练的团队,现在能做的只有点 star 等消息。\n\n对行业来说,这篇论文的方向感比单点跑分更值得关注:视频生成的下一个竞争点,可能不在更大的模型,而在谁能把 3D 几何变成训练数据与条件信号。法线转台这套做法成本不高(LoRA 级微调),却直接消掉了\"背面靠编\"的老毛病——当商品展示、数字资产这类应用开始要求 360 度一致性时,先解决几何的人先拿到门票。\n\n参考:arXiv:2610.35684(https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.35684);项目页 justin4ai.github.io\u002FOmniShowcase;GitHub github.com\u002Fjustin4ai\u002FOmniShowcase","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.35684","7437aeb9-930c-4866-a2e9-48003c1a792b",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"7b67033c-19e6-4052-a626-e681bba64c7a","diffusion",{"id":19,"name":20,"slug":20,"description":14,"color":14},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":22,"name":23,"slug":23,"description":14,"color":14},"ebe5dcd1-46b1-4298-b8c2-8e0e2f456e56","video-generation",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"4027eb7e-28a6-4d3c-b4d8-f956b6327412","en","OmniShowcase: 360° Subject Video From 4 Reference Views","OmniShowcase (KAIST AI) turns 4 reference views into subject-consistent 360° videos, halving geometric error vs multi-view baselines; code unreleased.","A full 360-degree orbit video of a product now takes just 4 reference images. On Oct 26, a team led by KAIST AI, together with Inha University and FLIPTION, posted OmniShowcase to arXiv (2610.35684); three days later it ranked #2 on Hugging Face Daily Papers for Oct 29 with 42 upvotes. The problem it targets is precise: in subject-to-video (S2V) generation, once the camera swings to surfaces the references never showed, the model starts inventing the subject.\n\n## One reference image cannot carry 360 degrees\n\nThe standard subject-to-video recipe conditions on a single reference image, forcing the video model to infer the subject's unseen appearance as the viewpoint changes. Recent methods extend this to multiple reference views, which adds information but never tells the model explicitly what 3D structure connects those observations. The paper points to a subtler flaw in training data itself: existing corpora do not guarantee that viewpoint changes preserve the same underlying 3D subject — the model is effectively studying a course with no answer key.\n\n## Two weapons: geometry-built data, normals as a turntable\n\nThe first weapon is a fully automated data synthesis pipeline. Starting from one curated 3D asset, Blender renders both the target orbit video and four reference views, and the target is composited into a prompt-driven scene. The key detail: only the background is inpainted, so the subject's far side always comes from real rendering and is never hallucinated. This yields Showcase-4K, a training corpus covering twelve camera motions across 88 trajectory configurations, about 4,000 videos, each with reconstruction-derived surface-normal conditioning. A held-out ShowcaseBench ships the source meshes and camera parameters behind every video — an evaluation set that carries its own answer key, which is rare in video generation.\n\nThe second weapon is the geometry conditioning: the four reference views are reconstructed in 3D and rendered as a turntable of surface-normal maps, then fed into the reference pathway alongside the references. The base model is a Wan2.1 image-to-video diffusion transformer with lightweight LoRA tuning; the normal turntable requires no frame-wise alignment with the generated video and no per-subject optimization. The turntable itself is a fixed yaw sweep, but the generated camera is not bound to it — the target follows whichever trajectory the prompt names, from 360-degree orbits and helix descents to trucks, arcs and dolly moves.\n\n## Numbers: geometric error halved, aesthetics not overclaimed\n\nOn its own ShowcaseBench, OmniShowcase scores DINO 0.799 and CLIP 0.925 for identity and GT-CD 0.0824 for geometry, leading all identity- and geometry-related metrics against four multi-view baselines: Phantom-MV, MAGREF-MV, MV-S2V and 3DreamBooth. Versus the strongest of them, geometric error drops about 52% on the held-out set and about 46% on the real-world NAVI objects — the \"halved\" phrasing comes from the authors' own paper-page note and checks out against the quantitative tables on the project page. Worth noting: its aesthetic score of 0.574 sits below MAGREF-MV's 0.614, so geometry is a hard win while visual appeal only ties. The ablation is equally clean: removing the normals, swapping in a textured render, or thinning the turntable degrades MEt3R from 0.190 to around 0.26, confirming the normal condition as the source of the gains.\n\n## Cold water: paper first, all four artifacts unreleased\n\nThe GitHub repository was created on Oct 29 with a single initial commit and 8 stars. On the README release checklist, all four items — Showcase-4K, ShowcaseBench, inference code, training code — remain unchecked, and the Hugging Face page shows zero models citing the paper. In other words, the paper and the demos come first; the full reproducible artifact set is still on the way. Teams waiting to train 3D consistency on the dataset can do little right now except star the repo.\n\nFor the industry, the direction matters more than any single benchmark: the next battleground in video generation may not be bigger models, but who can turn 3D geometry into training data and conditioning signals. The normal-turntable recipe is cheap — LoRA-level tuning — yet it directly kills the old \"hallucinate the far side\" failure. As product showcases and digital-asset workflows start demanding 360-degree consistency, whoever solves geometry first gets the ticket.\n\nReferences: arXiv:2610.35684 (https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.35684); project page justin4ai.github.io\u002FOmniShowcase; GitHub github.com\u002Fjustin4ai\u002FOmniShowcase","omnishowcase-360-subject-video","2026-10-29T23:06:48Z","2026-10-29T23:06:52.225673Z","2026-10-29T23:06:52.225683Z",true,"agent",348,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"03d88e2a-526c-4c39-b9b6-e90ac0ff65e0","PixelUMM 拆掉 VAE:NVIDIA 用 8B 单 Transformer 在像素空间统一图文视频","pixelumm-encoder-free-unified-multimodal-8b","2026-09-29T21:59:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"62b2e6d9-4ac7-457c-a30c-548c713730ad","PhyCo：让视频生成模型「理解」物理世界","phyco-cvpr-2026-physics-video-controlnet-vlm-reward","2026-05-01T16:00:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"85b85139-0d39-464a-99cf-415fd498e9d2","L3G-World:高斯记忆让世界模型9 FPS实时生长","l3g-world-gaussian-memory-realtime","2026-10-30T21:12:08+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"dc534f1a-dbe2-4f79-9557-bbb43a82f24f","视频模型总在绕过 MLLM,三段训练把它拉回来","fapl-mllm-latents-cinematic-video-generation","2026-10-23T09:20:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"af21d26d-d9cd-4526-ac4a-66366a45848c","AV-GRPO:8张A800给22B音视频模型做RL后训练","av-grpo-audio-video-diffusion-rl","2026-09-27T17:08:13+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"9210c86b-3e1d-434b-bbfd-78e62c698aed","WorldCrafter 开源:给视频世界模型装上可查询的 3D 记忆,转一圈回来还是那个房间","worldcrafter-video-world-model-3d-memory","2026-09-22T19:08:48+00:00"]