ViewMask-1-to-3: Multi-View Consistent Image Generation via Multimodal Discrete Diffusion Models

Motivated by discrete diffusion's success in language-vision modeling, we explore its potential for multi-view generation, a task dominated by continuous approaches. We introduce ViewMask-1-to-3, formulating multi-view generation as a discrete sequence modeling problem where each viewpoint is represented as visual tokens from MAGVIT-v2. Through discrete diffusion via masked token prediction, our approach enables progressive multi-view generation via iterative token unmasking, unifying language and vision in a shared token space. Importantly, simple random masking combined with self-attention naturally encourages cross-view consistency without specialized architectures or 3D geometric priors. Our method outperforms the baseline on the GSO and 3D-FUTURE benchmarks, ranking first on average across standard image metricsand, and achieving a 10.6% higher IoU than continuous diffusion models on 3D-FUTURE. Furthermore, the proposed framework can be naturally extended to support text-to-image generation and multimodal understanding, highlighting its potential toward a more unified paradigm for multimodal understanding and generation.

Generating multiple consistent views of a 3D object from a single image — like imagining what the back of a chair looks like from only a front photo — is a hard problem in computer vision. Most existing methods treat images as continuous pixel grids, limiting their flexibility and scalability. We represent images as discrete visual tokens, similar to words in a sentence, and generate new viewpoints by progressively revealing masked tokens. This simple design naturally keeps different views consistent without requiring any 3D geometric knowledge. Our method outperforms existing approaches on standard benchmarks, and the same framework extends naturally to text-to-image generation and visual understanding.