Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle
This addresses a key bottleneck in 3D content generation for applications like gaming and VR, though it is incremental as it builds on existing multi-view diffusion and reconstruction methods.
The paper tackles the problem of low-quality and inconsistent multi-view images in image-to-3D generation, which degrade 3D reconstruction, by proposing Cycle3D, a framework that cyclically integrates 2D diffusion and 3D reconstruction modules to achieve high-quality and consistent 3D content, as demonstrated by superior performance in experiments.
Recent 3D large reconstruction models typically employ a two-stage process, including first generate multi-view images by a multi-view diffusion model, and then utilize a feed-forward model to reconstruct images to 3D content.However, multi-view diffusion models often produce low-quality and inconsistent images, adversely affecting the quality of the final 3D reconstruction. To address this issue, we propose a unified 3D generation framework called Cycle3D, which cyclically utilizes a 2D diffusion-based generation module and a feed-forward 3D reconstruction module during the multi-step diffusion process. Concretely, 2D diffusion model is applied for generating high-quality texture, and the reconstruction model guarantees multi-view consistency.Moreover, 2D diffusion model can further control the generated content and inject reference-view information for unseen views, thereby enhancing the diversity and texture consistency of 3D generation during the denoising process. Extensive experiments demonstrate the superior ability of our method to create 3D content with high-quality and consistency compared with state-of-the-art baselines.