✨How Much 3D Do Video Foundation Models Encode?

📝 Summary:
A new framework quantifies 3D understanding in Video Foundation Models VidFMs. VidFMs, trained only on video, show strong 3D awareness, often surpassing expert 3D models, providing insights for 3D AI.

🔹 Publication Date: Published on Dec 23

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2512.19949
• PDF: https://arxiv.org/pdf/2512.19949
• Project Page: https://vidfm-3d-probe.github.io/
• Github: https://vidfm-3d-probe.github.io

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#VideoFoundationModels #3DUnderstanding #ComputerVision #AIResearch #DeepLearning

🔥 VLM3: Vision Language Models Are Native 3D Learners

💡 The paper VLM3 Vision Language Models Are Native 3D Learners presents a study that challenges the common approach to 3D understanding tasks in computer vision. Typically these tasks rely on specialized vision models with complex designs and extensive data augmentation. However the authors argue that vision language models can be adapted for 3D understanding tasks through simple architectural modifications and text-based training.

The problem addressed in this paper is that 3D understanding tasks such as depth estimation and object-level 3D understanding are currently dominated by expert vision models that have complex task-specific designs. The authors propose that vision language models can be native 3D learners and achieve comparable performance to these specialized models.

The method used in this study involves making three simple modifications to standard vision language models. These modifications include focal length unification, text-based pixel reference, and data mixture and scaling. The authors propose VLM3, a scalable method that enables standard vision language models to master diverse 3D tasks without requiring complex designs or extensive data augmentation.

The results of the study show that VLM3 advances the depth estimation accuracy of vision language models by a large margin, from 0.84 to 0.9. Additionally, VLM3 enables diverse 3D tasks such as pixel correspondence, camera pose estimation, and object-level 3D understanding, matching the accuracy of expert vision models while maintaining standard architectures and text-based training. Overall, the paper presents a new paradigm for simple and scalable 3D learning, demonstrating that vision language models can be effective native 3D learners.

📅 Published on May 28

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2605.30561
• PDF: https://arxiv.org/pdf/2605.30561

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#VisionLanguageModels #3DUnderstanding #DepthEstimation #ObjectLevel3D #ComputerVisionModels

🔥 Hunyuan3D-Buffalo 1.0: A Unified Multimodal Model for Scalable 3D Generation, Understanding, and Editing

💡 The paper introduces Hunyuan3D-Buffalo 1.0, a unified multimodal model that integrates 3D understanding, generation, editing, and part generation into a single architecture. The main problem addressed is the lack of large-scale and geometrically consistent editing data for 3D modeling, which hinders the development of unified 3D models. To overcome this limitation, the authors propose a framework that combines two main components: Hunyuan3D-VLM for semantic, structural, and spatial understanding, and Hunyuan3D DiT for high-fidelity 3D synthesis.

The method involves training the model on an 87M-scale 3D multimodal corpus, which includes 25M understanding samples, 50M text-to-3D pairs, and 12M editing pairs generated using Nano3D-v2. The model uses multimodal semantic conditions for generation and conditions the diffusion process on the source object representation for editing and part generation, preserving the overall structure and unedited regions.

The results show that Hunyuan3D-Buffalo 1.0 achieves state-of-the-art or leading performance on text-to-3D generation and 3D editing benchmarks, while also exhibiting strong understanding and part-generation capabilities. The analysis demonstrates that unified 3D multimodal training is effective, as both generation and understanding improve editing. The paper contributes to the development of scalable and unified 3D modeling by providing a framework that can perform multiple tasks within a single architecture, enabling more efficient and effective 3D modeling and editing.

📅 Published on Aug 3

🔗 Links:
• GitHub: https://github.com/huggingface
• arXiv: https://arxiv.org/abs/2608.02711
• PDF: https://arxiv.org/pdf/2608.02711
• Project Page: https://tencent-hunyuan.github.io/Hunyuan3D-Buffalo1.0/

━━━━━━━━━━━━━━━━━━━━━━━━
📢 By: https://xn--r1a.website/PaperNexus

#3DGeneration #MultimodalModeling #3DUnderstanding #GeometricEditing #3DModelingTechniques