VividDreamer: Invariant Score Distillation For Hyper-Realistic Text-to-3D Generation
This addresses a key bottleneck in generating hyper-realistic 3D objects from text, offering an incremental improvement over existing SDS methods.
The paper tackles over-saturation and over-smoothing problems in Score Distillation Sampling (SDS) for text-to-3D generation by introducing Invariant Score Distillation (ISD), which replaces the reconstruction term with an invariant score term, resulting in greatly enhanced realism through single-stage optimization.
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). In this paper, SDS is decoupled into a weighted sum of two components: the reconstruction term and the classifier-free guidance term. We experimentally found that over-saturation stems from the large classifier-free guidance scale and over-smoothing comes from the reconstruction term. To overcome these problems, ISD utilizes an invariant score term derived from DDIM sampling to replace the reconstruction term in SDS. This operation allows the utilization of a medium classifier-free guidance scale and mitigates the reconstruction-related errors, thus preventing the over-smoothing and over-saturation of results. Extensive experiments demonstrate that our method greatly enhances SDS and produces realistic 3D objects through single-stage optimization.