ViewFusion: Towards Multi-View Consistency via Interpolated Denoising

Xianghui Yang, Yan Zuo, Sameera Ramasinghe, Loris Bazzani, Gil Avraham, Anton van den Hengel; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 9870-9880

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.