Score Distillation via Reparametrized DDIM

• URL: http://arxiv.org/abs/2405.15891v3
• Date: Thu, 10 Oct 2024 09:25:13 GMT
• Title: Score Distillation via Reparametrized DDIM
• Authors: Artem Lukoianov, Haitz Sáez de Ocáriz Borde, Kristjan Greenewald, Vitor Campagnolo Guizilini, Timur Bagautdinov, Vincent Sitzmann, Justin Solomon,
• Abstract summary: We show that the image guidance used in Score Distillation Sampling can be understood as the velocity field of a 2D denoising generative process. We show that a better noise approximation can be recovered by inverting DDIM in each SDS update step. Our method achieves better or similar 3D generation quality compared to other state-of-the-art Score Distillation methods.
• Score: 14.754513907729878
• License:
• Abstract: While 2D diffusion models generate realistic, high-detail images, 3D shape generation methods like Score Distillation Sampling (SDS) built on these 2D diffusion models produce cartoon-like, over-smoothed shapes. To help explain this discrepancy, we show that the image guidance used in Score Distillation can be understood as the velocity field of a 2D denoising generative process, up to the choice of a noise term. In particular, after a change of variables, SDS resembles a high-variance version of Denoising Diffusion Implicit Models (DDIM) with a differently-sampled noise term: SDS introduces noise i.i.d. randomly at each step, while DDIM infers it from the previous noise predictions. This excessive variance can lead to over-smoothing and unrealistic outputs. We show that a better noise approximation can be recovered by inverting DDIM in each SDS update step. This modification makes SDS's generative process for 2D images almost identical to DDIM. In 3D, it removes over-smoothing, preserves higher-frequency detail, and brings the generation quality closer to that of 2D samplers. Experimentally, our method achieves better or similar 3D generation quality compared to other state-of-the-art Score Distillation methods, all without training additional neural networks or multi-view supervision, and providing useful insights into relationship between 2D and 3D asset generation with diffusion models.

Related papers

• VividDreamer: Invariant Score Distillation For Hyper-Realistic Text-to-3D Generation [33.05759961083337]
  This paper presents Invariant Score Distillation (ISD), a novel method for high-fidelity text-to-3D generation. ISD aims to tackle the over-saturation and over-smoothing problems in Score Distillation Sampling (SDS)
  arXiv  Detail & Related papers  (2024-07-13T09:33:16Z)
• VividDreamer: Towards High-Fidelity and Efficient Text-to-3D Generation [69.68568248073747]
  We propose Pose-dependent Consistency Distillation Sampling (PCDS), a novel yet efficient objective for diffusion-based 3D generation tasks. PCDS builds the pose-dependent consistency function within diffusion trajectories, allowing to approximate true gradients through minimal sampling steps. For efficient generation, we propose a coarse-to-fine optimization strategy, which first utilizes 1-step PCDS to create the basic structure of 3D objects, and then gradually increases PCDS steps to generate fine-grained details.
  arXiv  Detail & Related papers  (2024-06-21T08:21:52Z)
• ExactDreamer: High-Fidelity Text-to-3D Content Creation via Exact Score Matching [10.362259643427526]
  Current approaches often adapt pre-trained 2D diffusion models for 3D synthesis. Over-smoothing poses a significant limitation on the high-fidelity generation of 3D models. LucidDreamer replaces the Denoising Diffusion Probabilistic Model (DDPM) in SDS with the Denoising Diffusion Implicit Model (DDIM)
  arXiv  Detail & Related papers  (2024-05-24T20:19:45Z)
• Flow Score Distillation for Diverse Text-to-3D Generation [23.38418695449777]
  Flow Score Distillation (FSD) substantially enhances generation diversity without compromising quality. Our validation experiments across various text-to-image Diffusion Models demonstrate that FSD substantially enhances generation diversity without compromising quality.
  arXiv  Detail & Related papers  (2024-05-16T06:05:16Z)
• Diffusion Time-step Curriculum for One Image to 3D Generation [91.07638345953016]
  Score distillation sampling(SDS) has been widely adopted to overcome the absence of unseen views in reconstructing 3D objects from a textbfsingle image. We find out the crux is the overlooked indiscriminate treatment of diffusion time-steps during optimization. We propose the Diffusion Time-step Curriculum one-image-to-3D pipeline (DTC123), which involves both the teacher and student models collaborating with the time-step curriculum in a coarse-to-fine manner.
  arXiv  Detail & Related papers  (2024-04-06T09:03:18Z)
• Consistent3D: Towards Consistent High-Fidelity Text-to-3D Generation with Deterministic Sampling Prior [87.55592645191122]
  Score distillation sampling (SDS) and its variants have greatly boosted the development of text-to-3D generation, but are vulnerable to geometry collapse and poor textures yet. We propose a novel and effective "Consistent3D" method that explores the ODE deterministic sampling prior for text-to-3D generation. Experimental results show the efficacy of our Consistent3D in generating high-fidelity and diverse 3D objects and large-scale scenes.
  arXiv  Detail & Related papers  (2024-01-17T08:32:07Z)
• NeuSD: Surface Completion with Multi-View Text-to-Image Diffusion [56.98287481620215]
  We present a novel method for 3D surface reconstruction from multiple images where only a part of the object of interest is captured. Our approach builds on two recent developments: surface reconstruction using neural radiance fields for the reconstruction of the visible parts of the surface, and guidance of pre-trained 2D diffusion models in the form of Score Distillation Sampling (SDS) to complete the shape in unobserved regions in a plausible manner.
  arXiv  Detail & Related papers  (2023-12-07T19:30:55Z)
• StableDreamer: Taming Noisy Score Distillation Sampling for Text-to-3D [88.66678730537777]
  We present StableDreamer, a methodology incorporating three advances. First, we formalize the equivalence of the SDS generative prior and a simple supervised L2 reconstruction loss. Second, our analysis shows that while image-space diffusion contributes to geometric precision, latent-space diffusion is crucial for vivid color rendition.
  arXiv  Detail & Related papers  (2023-12-02T02:27:58Z)
• HD-Fusion: Detailed Text-to-3D Generation Leveraging Multiple Noise Estimation [43.83459204345063]
  We propose a novel approach that combines multiple noise estimation processes with a pretrained 2D diffusion prior. Results show that the proposed approach can generate high-quality details compared to the baselines.
  arXiv  Detail & Related papers  (2023-07-30T09:46:22Z)
• Improving 3D Imaging with Pre-Trained Perpendicular 2D Diffusion Models [52.529394863331326]
  We propose a novel approach using two perpendicular pre-trained 2D diffusion models to solve the 3D inverse problem. Our method is highly effective for 3D medical image reconstruction tasks, including MRI Z-axis super-resolution, compressed sensing MRI, and sparse-view CT.
  arXiv  Detail & Related papers  (2023-03-15T08:28:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.