~ similar to 2609.31595· 20 results
Shaohui Dai, Yansong Qu, You Shen, Shengchuan Zhang +1 more
The paper introduces PAR3D, a unified part-aware 3D-MLLM framework, to enhance 3D scene understanding by enabling models to reason about and ground both whole objects and their fine-grained parts.
Yikang Yang, Zhanpeng Hu, Youtian Lin, Mengqi Zhou +4 more
P3D-Bench benchmark for parametric 3D generation
Zhipeng Cai, Zhuang Liu, Yunyang Xiong, Zechun Liu +2 more
The paper proposes VLM3, a simple, scalable method that demonstrates standard Vision Language Models (VLMs) can natively learn 3D understanding by focusing on architectural simplicity and specific dat…
Jiangwei Ren, Xingyu Jiang, Zijie Song, Wei Xu +3 more
This paper proposes Wat3R, a cross-domain semi-supervised learning framework for adapting 3D reconstruction models from air to underwater scenes using unlabeled real underwater video footage and a tea…
Tianjiao Yu, Xinzhuo Li, Yifan Shen, Ying Shen +3 more
This paper introduces SILSA, a topology-aware 3D generation framework that improves structural fidelity and reduces generation cost.
Yuming Zhao, Junhui Hou, Qijian Zhang, Jia Qin +1 more
The paper introduces PRISM, a novel representation learning framework that learns isometric embeddings by explicitly modeling the intrinsic geodesic metric of 3D surfaces, achieving superior performan…
Jiawei Li, Ziyi Liu, Weijie Shi, Long Chen +2 more
SSR3D-LLM introduces a structured spatial reasoning interface for unified 3D-LLMs, allowing fine-grained object grounding by generating and processing sequential latent spatial steps.
Wenhao Li, Xueying Jiang, Quanhao Qian, Deli Zhao +3 more
The paper introduces VLM-IE3D, a framework that enhances 2D vision-language models with implicit and explicit 3D geometries learned from RGB videos, achieving superior performance on various 3D tasks.
The paper proposes AlignG, a method that learns context-conditioned predicate semantics by using prototype feedback to adapt relation representations based on image-specific evidence, significantly im…
Antoine Guédon, Shu Nakamura, Nicolas Dufour, Jiahui Lei +2 more
Surflo compresses a variable number of RGB views into a single global latent state and decodes oriented 3D surface points.
Chun-Hsiao Yeh, Shengyi Qian, Manchen Wang, Yi Ma +2 more
The paper proposes GASP, a framework that injects fundamental geometric priors directly into Vision-Language Models (VLMs) using ground-truth video geometry, significantly enhancing 3D spatial reasoni…
Yipeng Gao, Lei Shu, Genzhi Ye, Xi Xiong +4 more
The paper introduces 3DCodeBench, a systematic benchmark and platform for evaluating Vision-Language Model (VLM) agents' ability to generate procedural 3D models from text and images using code.
The paper introduces Staged Executable Inverse Graphics (SEIG), an agentic framework that uses general-purpose Vision-Language Models (VLMs) to reconstruct editable 3D scenes directly into executable…
Shuo Lu, Yinuo Xu, Kecheng Yu, Siru Jiang +7 more
The paper introduces WorldCoder-Bench, a comprehensive benchmark and evaluation protocol for testing LLMs' ability to autonomously generate complex, physically grounded, and interactive 3D web worlds.
Maëlic Neau, Salim Baloch, Jakob Suchan, Zoe Falomir +1 more
The paper proposes a model-agnostic framework to improve scene graph generation models under annotation sparsity by capturing visual commonsense knowledge.
Reasmory introduces a structured programming framework that uses explicit 3D memory and a Domain-Specific Language (DSL) to reliably enhance Vision-Language Models' spatial reasoning capabilities, ach…
Zhenxiang Ma, Zeyu He, Yuanzhen Zhou, Zhenyu Yang +5 more
The paper presents Odin, a distributed training system for point-based neural rendering (PBNR) that replaces global barriers with primitive-level synchronization, improving throughput and hiding criti…
The paper introduces S2MDF, a plug-and-play module that enforces a hard constraint to eliminate interpenetrations in multi-object Signed Distance Field (SDF) representations, significantly improving p…
Sayan Paul, Sourav Ghosh, Siddharth Katageri, Soumyadip Maity +2 more
City-Mesh3R is a scalable, end-to-end framework that reconstructs high-fidelity, watertight 3D surface meshes of entire city-scale environments directly from large collections of multi-view images.