ViewFusion: Towards Multi-View Consistency via Interpolated Denoising

SeanYoung, Yan Zuo, Sameera Ramasinghe, Loris Bazzani, Gil Avraham, Anton van den Hengel

ViewFusion: Towards Multi-View Consistency via Interpolated Denoising: 14 upvotes on Hugging Face Daily Papers, #11 of 11 papers on 2024-03-01. Day-by-day upvote history.

Novel-view synthesis through diffusion models has demonstrated remarkable potential for generating diverse and high-quality images. Yet, the independent process of image generation in these prevailing methods leads to challenges in maintaining multiple-view consistency. To address this, we introduce ViewFusion, a novel, training-free algorithm that can be seamlessly integrated into existing pre-trained diffusion models. Our approach adopts an auto-regressive method that implicitly leverages previously generated views as context for the next view generation, ensuring robust multi-view consistency during the novel-view generation process. Through a diffusion process that fuses known-view information via interpolated denoising, our framework successfully extends single-view conditioned models to work in multiple-view conditional settings without any additional fine-tuning. Extensive experimental results demonstrate the effectiveness of ViewFusion in generating consistent and detailed novel views.

Paper page on Hugging Face · arXiv

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.