Saved in:
Bibliographic Details
Main Authors: Zhu, Ruishu, Huang, Zhihao, Sun, Jiacheng, Luo, Ping, Zhang, Hongyuan, Li, Xuelong
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.14099
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918385806213120
author Zhu, Ruishu
Huang, Zhihao
Sun, Jiacheng
Luo, Ping
Zhang, Hongyuan
Li, Xuelong
author_facet Zhu, Ruishu
Huang, Zhihao
Sun, Jiacheng
Luo, Ping
Zhang, Hongyuan
Li, Xuelong
contents 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 synthesis as a discrete sequence modeling problem where each viewpoint is represented as visual tokens from MAGVIT-v2. Through 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 metrics and improving IoU by 10.6% on 3D-FUTURE. This validates discrete diffusion as a promising candidate for multi-view generation.
format Preprint
id arxiv_https___arxiv_org_abs_2512_14099
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ViewMask-1-to-3: Multi-View Consistent Image Generation via Multimodal Diffusion Models
Zhu, Ruishu
Huang, Zhihao
Sun, Jiacheng
Luo, Ping
Zhang, Hongyuan
Li, Xuelong
Computer Vision and Pattern Recognition
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 synthesis as a discrete sequence modeling problem where each viewpoint is represented as visual tokens from MAGVIT-v2. Through 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 metrics and improving IoU by 10.6% on 3D-FUTURE. This validates discrete diffusion as a promising candidate for multi-view generation.
title ViewMask-1-to-3: Multi-View Consistent Image Generation via Multimodal Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.14099