Thinking with Camera: A Unified Multimodal Model for Camera-Centric Understanding and Generation
This work addresses the challenge of spatial intelligence in multimodal AI by enabling flexible camera-centric tasks, though it appears incremental as it builds on existing methods like diffusion and language models.
The paper tackles the problem of isolated camera-centric understanding and generation by presenting Puffin, a unified multimodal model that integrates language regression and diffusion-based generation to interpret and create scenes from arbitrary viewpoints, achieving superior performance over specialized models. It introduces a novel paradigm treating camera as language and is trained on a large-scale dataset of 4 million vision-language-camera triplets.
Camera-centric understanding and generation are two cornerstones of spatial intelligence, yet they are typically studied in isolation. We present Puffin, a unified camera-centric multimodal model that extends spatial awareness along the camera dimension. Puffin integrates language regression and diffusion-based generation to interpret and create scenes from arbitrary viewpoints. To bridge the modality gap between cameras and vision-language, we introduce a novel paradigm that treats camera as language, enabling thinking with camera. This guides the model to align spatially grounded visual cues with photographic terminology while reasoning across geometric context. Puffin is trained on Puffin-4M, a large-scale dataset of 4 million vision-language-camera triplets. We incorporate both global camera parameters and pixel-wise camera maps, yielding flexible and reliable spatial generation. Experiments demonstrate Puffin superior performance over specialized models for camera-centric generation and understanding. With instruction tuning, Puffin generalizes to diverse cross-view tasks such as spatial imagination, world exploration, and photography guidance. We will release the code, models, dataset pipeline, and benchmark to advance multimodal spatial intelligence research.