3D 视觉 - arXivDaily 专题

2311.15980 2024-03-22 cs.CV 74%

Direct2.5: Diverse Text-to-3D Generation via Multi-view 2.5D Diffusion

Yuanxun Lu, Jingyang Zhang, Shiwei Li, Tian Fang, David McKinnon, Yanghai Tsin, Long Quan, Xun Cao, Yao Yao

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments CVPR 2024 camera ready, including more evaluations and discussions. Project webpage: https://nju-3dv.github.io/projects/direct25

2402.00225 2024-02-05 cs.CV 74%

Geometry aware 3D generation from in-the-wild images in ImageNet

Qijia Shen, Guangrun Wang

专题命中 3D生成：3D generation(title)；分类 cs.CV

2310.12474 2024-01-19 cs.CV 74%

Enhancing High-Resolution 3D Generation through Pixel-wise Gradient Clipping

Zijie Pan, Jiachen Lu, Xiatian Zhu, Li Zhang

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments Accepted at ICLR 2024. Project page: https://fudan-zvg.github.io/PGC-3D

2401.07727 2024-01-17 cs.CV 74%

HexaGen3D: StableDiffusion is just one step away from Fast and Diverse Text-to-3D Generation

Antoine Mercier, Ramin Nakhli, Mahesh Reddy, Rajeev Yasarla, Hong Cai, Fatih Porikli, Guillaume Berger

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments 9 pages, 8 figures, 2 tables

2401.04092 2024-01-11 cs.CV 74%

GPT-4V(ision) is a Human-Aligned Evaluator for Text-to-3D Generation

Tong Wu, Guandao Yang, Zhibing Li, Kai Zhang, Ziwei Liu, Leonidas Guibas, Dahua Lin, Gordon Wetzstein

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments Project page: https://gpteval3d.github.io/ ; Code: https://github.com/3DTopia/GPTEval3D

2312.02201 2023-12-06 cs.CV 74%

ImageDream: Image-Prompt Multi-view Diffusion for 3D Generation

Peng Wang, Yichun Shi

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments project page: https://Image-Dream.github.io

2311.14494 2023-11-29 cs.CV 74%

MVControl: Adding Conditional Control to Multi-view Diffusion for Controllable Text-to-3D Generation

Zhiqi Li, Yiming Chen, Lingzhe Zhao, Peidong Liu

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments Project page: https://lizhiqi49.github.io/MVControl/

2310.16818 2023-10-27 cs.CV cs.CG 74%

DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior

Jingxiang Sun, Bo Zhang, Ruizhi Shao, Lizhen Wang, Wen Liu, Zhenda Xie, Yebin Liu

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments Project Page: https://mrtornado24.github.io/DreamCraft3D/

2308.11473 2023-08-23 cs.CV cs.AI 74%

IT3D: Improved Text-to-3D Generation with Explicit View Synthesis

Yiwen Chen, Chi Zhang, Xiaofeng Yang, Zhongang Cai, Gang Yu, Lei Yang, Guosheng Lin

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments Project Page: https://github.com/buaacyw/IT3D-text-to-3D

2307.16183 2023-08-01 cs.CV 74%

HD-Fusion: Detailed Text-to-3D Generation Leveraging Multiple Noise Estimation

Jinbo Wu, Xiaobo Gao, Xing Liu, Zhengyang Shen, Chen Zhao, Haocheng Feng, Jingtuo Liu, Errui Ding

专题命中 3D生成：3D generation(title)；分类 cs.CV

2212.00774 2022-12-02 cs.CV cs.LG 74%

Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation

Haochen Wang, Xiaodan Du, Jiahao Li, Raymond A. Yeh, Greg Shakhnarovich

专题命中 3D生成：3D generation(title)；分类 cs.CV

Comments project page https://pals.ttic.edu/p/score-jacobian-chaining

2607.16352 2026-07-21 cs.CV cs.AI cs.CL cs.GR cs.MA 新提交 73%

Clarify Before Executing: A Self-Evolving Agent for Resolving Intent Asymmetry in 3D Tool Orchestration

执行前澄清：用于解决3D工具编排中意图不对称的自进化智能体

Xiaoye Zhu, Weixin Li, Junan Huo, Bozhong Wang, Jia Zeng, Yi Yang, Cen Chen, Qi Liu

专题命中 3D生成：point cloud(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

AI总结研究3D资产创建中用户指令模糊与工具要求精确的意图不对称问题，提出自进化智能体CLARE，通过解耦生成管道、模拟多轮交互自我进化澄清策略，在新基准测试中取得领先性能，证明主动澄清对3D执行的关键作用。

Comments Accepted to ACM Multimedia 2026 (ACM MM 2026). 26 pages including appendix, 8 figures. Code: https://github.com/xyzhu1225/CLARE

2606.11446 2026-06-11 cs.CV cs.GR 新提交 73%

3D-CBM: A Framework for Concept-Based Interpretability in Generative 3D Modeling

3D-CBM：生成式3D建模中基于概念可解释性的框架

Ahmad Al-Kabbany

机构 * Yubree Labs ； Multimedia Interaction and Communication Lab, Arab Academy for Science and Technology（阿拉伯科学技术学院多媒体交互与通信实验室）

专题命中 3D生成：point cloud(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

AI总结提出将概念瓶颈模型（CBM）融入3D生成架构，通过多层级可解释原语和功能属性映射，实现语义可操控的3D生成，实验验证了高概念预测精度和交互式纠错能力。

2605.16355 2026-05-19 cs.GR cs.CV 73%

Generative 3D Gaussians with Learned Density Control

具有学习密度控制的生成3D高斯

Runjie Yan, Yan-Pei Cao, Peng Wang, Ding Liang, Yuan-Chen Guo

机构 * Institute for Interdisciplinary Information Sciences, Tsinghua University（清华大学交叉信息研究院）

专题命中 3D生成：Gaussian Splatting(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

AI总结本文提出Density-Sampled Gaussians (DeG)，通过学习概率密度函数实现3D表示的自适应密度控制，结合渲染监督优化空间密度和高斯属性，提升单图到3D生成的质量。

Comments 19 pages, 16 figures, SIGGRAPH Conference Papers '26

2603.29592 2026-04-01 cs.GR cs.CV 73%

Bioinspired123D: Generative 3D Modeling System for Bioinspired Structures

生物启发123D：用于生物启发结构的生成3D建模系统

Rachel K. Luu, Markus J. Buehler

专题命中 3D生成：point cloud(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

AI总结本文提出Bioinspired123D，一种轻量级模块化代码-几何管道，通过参数化程序直接生成可制造的3D结构，而非密集视觉表示。系统基于生物启发3D模型，结合多模态检索增强生成和视觉-语言模型批评者，实现了高效的文本到3D生成。

2503.12052 2026-04-01 cs.CV cs.GR 73%

A Text-to-3D Framework for Joint Generation of CG-Ready Humans and Compatible Garments

为生成兼容服装的CG-ready人类和服装联合生成的文本到3D框架

Zhiyao Sun, Yu-Hui Wen, Ho-Jui Fang, Sheng Ye, Matthieu Lin, Tian Lv, Yong-Jin Liu

专题命中 3D生成：NeRF(abstract)；3DGS(abstract)；分类 cs.CV、cs.GR

AI总结本文提出Tailor框架，通过语义解析、几何感知服装生成和一致纹理合成，实现高保真定制3D人像和物理兼容服装，优于现有方法。

Comments Project page: https://human-tailor.github.io

2506.15684 2025-06-19 cs.GR cs.CV cs.LG 73%

Nabla-R2D3: Effective and Efficient 3D Diffusion Alignment with 2D Rewards

Qingming Liu, Zhen Liu, Dinghuai Zhang, Kui Jia

机构 * The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））； Microsoft Research（微软研究院）

专题命中 3D生成：3D vision(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

Comments Technical Report (21 pages, 21 figures)

2312.09222 2024-04-25 cs.CV cs.GR 73%

Mosaic-SDF for 3D Generative Models

Lior Yariv, Omri Puny, Natalia Neverova, Oran Gafni, Yaron Lipman

专题命中 3D生成：point cloud(abstract)；3D generation(abstract)；分类 cs.CV、cs.GR

Comments More results and details can be found at https://lioryariv.github.io/msdf

2606.30429 2026-07-01 cs.LG 版本更新 71%

Arko-T: A Foundation Model for Text-to-Structured 3D Generation

Arko-T: 面向文本到结构化3D生成的基础模型

Liang Wang, Zhaoyang Xi, Zekai Xiang, Heng Meng, Qishan Zhang, Pingyi Zhou, Jin Liu, Litao Chen

机构 * Spatial Design Intelligence Lab, BitInf Ltd.（BitInf有限公司空间设计智能实验室）； School of Computer Science, Wuhan University（武汉大学计算机学院）； College of Computer and Information Engineering, Nanjing Tech University（南京工业大学计算机与信息工程学院）

专题命中 3D生成：3D generation(title)

AI总结提出4B参数文本到设计模型Arko-T，通过设计状态对齐实现从自然语言到可编辑参数化CAD程序的生成，在12项指标中8项最优，成本仅为前沿模型的十分之一。

2606.20364 2026-06-19 cs.LG 新提交 71%

Judging to Improve: A De-biased VLM-as-3D-Judge Protocol for Single-Image 3D Generation

评判以改进：一种去偏的 VLM-as-3D-Judge 协议用于单图像 3D 生成

Ali Asaria, Tony Salomone, Deep Gandhi

机构 * Transformer Lab

专题命中 3D生成：3D generation(title)

AI总结本文提出一种去偏的跨模型 VLM-as-3D-Judge 协议，将评判者从排序扩展到优化，通过训练与评估评判者分离、位置偏差校正及修复三种失效模式，实现轻量级适应下与强基线的匹配。

2509.07435 2026-04-07 cs.CV 70%

DreamLifting: A Plug-in Module Lifting MV Diffusion Models for 3D Asset Generation

DreamLifting：一种插件模块用于提升MV扩散模型以生成3D资产

Ze-Xin Yin, Jiaxiong Qiu, Liu Liu, Xinjie Wang, Wei Sui, Zhizhong Su, Jian Yang, Jin Xie

机构 * PCA Lab, VCIP, College of Computer Science, Nankai University（南开大学计算机学院VCIP实验室PCA Lab）； School of Intelligence Science and Technology, Nanjing University（南京大学智能科学与技术学院）； Horizon Robotics（地平线机器人）； D-Robotics

专题命中 3D生成：Gaussian Splatting(abstract)；3D generation(abstract)；分类 cs.CV

AI总结本文提出LGAA框架，通过多视角扩散先验统一几何与PBR材质建模，结合模块化设计和改进的VAE实现高效3D资产生成，实验表明其在文本和图像条件下均表现优异。

Comments 16 pages, 9 figures, TVCG 2026, project page: https://zx-yin.github.io/dreamlifting/

Journal ref IEEE Transactions on Visualization and Computer Graphics 2026

2508.11952 2026-03-10 cs.CV 70%

UniUGG: Unified 3D Understanding and Generation via Geometric-Semantic Encoding

UniUGG: 通过几何-语义编码实现统一的3D理解与生成

Yueming Xu, Jiahui Zhang, Ze Huang, Yurui Chen, Yanpeng Zhou, Zhenyu Chen, Yu-Jie Yuan, Pengxiang Xia, Guowei Huang, Xinyue Cai, Zhongang Qi, Xingyue Quan, Jianye Hao, Hang Xu, Li Zhang

机构 * Fudan University（复旦大学）； Huawei Noah’s Ark Lab（华为诺亚实验室）

专题命中 3D生成：3D generation(abstract)；spatial understanding(abstract)；分类 cs.CV

AI总结 UniUGG通过几何-语义编码实现3D理解和生成的统一框架，结合大语言模型和潜在扩散模型，提升空间理解和生成能力。

2508.09973 2025-08-14 cs.CV 70%

PERSONA: Personalized Whole-Body 3D Avatar with Pose-Driven Deformations from a Single Image

Geonhee Sim, Gyeongsik Moon

机构 * Dept. of CSE, Korea University（计算机科学与工程系，韩国大学）

专题命中 3D生成：NeRF(abstract)；3DGS(abstract)；分类 cs.CV

Comments Accepted to ICCV 2025. https://mks0601.github.io/PERSONA/

2504.15835 2025-04-23 cs.CV 70%

Text-based Animatable 3D Avatars with Morphable Model Alignment

Yiqian Wu, Malte Prinzler, Xiaogang Jin, Siyu Tang

机构 * ETH Zürich（苏黎世联邦理工学院）； State Key Lab of CAD&CG, Zhejiang University（浙江大学计算机辅助设计与图形学国家重点实验室）

专题命中 3D生成：3DGS(abstract)；3D generation(abstract)；分类 cs.CV

2410.20723 2024-10-29 cs.CV 70%

CompGS: Unleashing 2D Compositionality for Compositional Text-to-3D via Dynamically Optimizing 3D Gaussians

Chongjian Ge, Chenfeng Xu, Yuanfeng Ji, Chensheng Peng, Masayoshi Tomizuka, Ping Luo, Mingyu Ding, Varun Jampani, Wei Zhan

专题命中 3D生成：Gaussian Splatting(abstract)；3D generation(abstract)；分类 cs.CV

2406.07742 2024-06-13 cs.CV 70%

C3DAG: Controlled 3D Animal Generation using 3D pose guidance

Sandeep Mishra, Oindrila Saha, Alan C. Bovik

专题命中 3D生成：NeRF(abstract)；3D generation(abstract)；分类 cs.CV

2401.10215 2024-01-19 cs.CV 70%

GPAvatar: Generalizable and Precise Head Avatar from Image(s)

Xuangeng Chu, Yu Li, Ailing Zeng, Tianyu Yang, Lijian Lin, Yunfei Liu, Tatsuya Harada

专题命中 3D生成：point cloud(abstract)；novel view synthesis(abstract)；分类 cs.CV

Comments ICLR 2024, code is available at https://github.com/xg-chu/GPAvatar

2311.13617 2023-11-27 cs.CV 70%

Boosting3D: High-Fidelity Image-to-3D by Boosting 2D Diffusion Prior to 3D Prior with Progressive Learning

Kai Yu, Jinlin Liu, Mengyang Feng, Miaomiao Cui, Xuansong Xie

专题命中 3D生成：NeRF(abstract)；3D generation(abstract)；分类 cs.CV

Comments 8 pages, 7 figures, 1 table

2311.09217 2023-11-16 cs.CV 70%

DMV3D: Denoising Multi-View Diffusion using 3D Large Reconstruction Model

Yinghao Xu, Hao Tan, Fujun Luan, Sai Bi, Peng Wang, Jiahao Li, Zifan Shi, Kalyan Sunkavalli, Gordon Wetzstein, Zexiang Xu, Kai Zhang

专题命中 3D生成：NeRF(abstract)；3D generation(abstract)；分类 cs.CV

Comments Project Page: https://justimyhxu.github.io/projects/dmv3d/

2603.23386 2026-03-25 cs.CV cs.GR cs.RO 67%

SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets via MLLM

SIMART: 通过MLLM将单体网格分解为可模拟的关节化资产

Chuanrui Zhang, Minghan Qin, Yuang Wang, Baifeng Xie, Hang Li, Ziwei Wang

机构 * ByteDance Seed（字节跳动种子）； NTU（国立台湾大学）

专题命中 3D生成：3D generation(abstract)；分类 cs.CV、cs.GR、cs.RO

AI总结 SIMART通过统一的MLLM框架实现部分级分解和运动学预测，减少3D令牌数量，提升多部分组装的保真度，并在PartNet-Mobility和野外AIGC数据集上取得最佳性能，支持物理仿真的机器人模拟。

1. 3D生成 447 篇

Direct2.5: Diverse Text-to-3D Generation via Multi-view 2.5D Diffusion

Geometry aware 3D generation from in-the-wild images in ImageNet

Enhancing High-Resolution 3D Generation through Pixel-wise Gradient Clipping

HexaGen3D: StableDiffusion is just one step away from Fast and Diverse Text-to-3D Generation

GPT-4V(ision) is a Human-Aligned Evaluator for Text-to-3D Generation

ImageDream: Image-Prompt Multi-view Diffusion for 3D Generation

MVControl: Adding Conditional Control to Multi-view Diffusion for Controllable Text-to-3D Generation

DreamCraft3D: Hierarchical 3D Generation with Bootstrapped Diffusion Prior

IT3D: Improved Text-to-3D Generation with Explicit View Synthesis

HD-Fusion: Detailed Text-to-3D Generation Leveraging Multiple Noise Estimation

Score Jacobian Chaining: Lifting Pretrained 2D Diffusion Models for 3D Generation

Clarify Before Executing: A Self-Evolving Agent for Resolving Intent Asymmetry in 3D Tool Orchestration

3D-CBM: A Framework for Concept-Based Interpretability in Generative 3D Modeling

Generative 3D Gaussians with Learned Density Control

Bioinspired123D: Generative 3D Modeling System for Bioinspired Structures

A Text-to-3D Framework for Joint Generation of CG-Ready Humans and Compatible Garments

Nabla-R2D3: Effective and Efficient 3D Diffusion Alignment with 2D Rewards

Mosaic-SDF for 3D Generative Models

Arko-T: A Foundation Model for Text-to-Structured 3D Generation

Judging to Improve: A De-biased VLM-as-3D-Judge Protocol for Single-Image 3D Generation

DreamLifting: A Plug-in Module Lifting MV Diffusion Models for 3D Asset Generation

UniUGG: Unified 3D Understanding and Generation via Geometric-Semantic Encoding

PERSONA: Personalized Whole-Body 3D Avatar with Pose-Driven Deformations from a Single Image

Text-based Animatable 3D Avatars with Morphable Model Alignment

CompGS: Unleashing 2D Compositionality for Compositional Text-to-3D via Dynamically Optimizing 3D Gaussians

C3DAG: Controlled 3D Animal Generation using 3D pose guidance

GPAvatar: Generalizable and Precise Head Avatar from Image(s)

Boosting3D: High-Fidelity Image-to-3D by Boosting 2D Diffusion Prior to 3D Prior with Progressive Learning

DMV3D: Denoising Multi-View Diffusion using 3D Large Reconstruction Model

SIMART: Decomposing Monolithic Meshes into Sim-ready Articulated Assets via MLLM