ModeDreamer: Mode Guiding Score Distillation for Text-to-3D Generation using Reference Image Prompts

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Tran, Uy Dieu, Luu, Minh, Nguyen, Phong Ha, Nguyen, Khoi, Hua, Binh-Son
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910855204962304
author Tran, Uy Dieu
Luu, Minh
Nguyen, Phong Ha
Nguyen, Khoi
Hua, Binh-Son
author_facet Tran, Uy Dieu
Luu, Minh
Nguyen, Phong Ha
Nguyen, Khoi
Hua, Binh-Son
contents Existing Score Distillation Sampling (SDS)-based methods have driven significant progress in text-to-3D generation. However, 3D models produced by SDS-based methods tend to exhibit over-smoothing and low-quality outputs. These issues arise from the mode-seeking behavior of current methods, where the scores used to update the model oscillate between multiple modes, resulting in unstable optimization and diminished output quality. To address this problem, we introduce a novel image prompt score distillation loss named ISD, which employs a reference image to direct text-to-3D optimization toward a specific mode. Our ISD loss can be implemented by using IP-Adapter, a lightweight adapter for integrating image prompt capability to a text-to-image diffusion model, as a mode-selection module. A variant of this adapter, when not being prompted by a reference image, can serve as an efficient control variate to reduce variance in score estimates, thereby enhancing both output quality and optimization stability. Our experiments demonstrate that the ISD loss consistently achieves visually coherent, high-quality outputs and improves optimization speed compared to prior text-to-3D methods, as demonstrated through both qualitative and quantitative evaluations on the T3Bench benchmark suite.
format Preprint
id arxiv_https___arxiv_org_abs_2411_18135
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ModeDreamer: Mode Guiding Score Distillation for Text-to-3D Generation using Reference Image Prompts
Tran, Uy Dieu
Luu, Minh
Nguyen, Phong Ha
Nguyen, Khoi
Hua, Binh-Son
Computer Vision and Pattern Recognition
Existing Score Distillation Sampling (SDS)-based methods have driven significant progress in text-to-3D generation. However, 3D models produced by SDS-based methods tend to exhibit over-smoothing and low-quality outputs. These issues arise from the mode-seeking behavior of current methods, where the scores used to update the model oscillate between multiple modes, resulting in unstable optimization and diminished output quality. To address this problem, we introduce a novel image prompt score distillation loss named ISD, which employs a reference image to direct text-to-3D optimization toward a specific mode. Our ISD loss can be implemented by using IP-Adapter, a lightweight adapter for integrating image prompt capability to a text-to-image diffusion model, as a mode-selection module. A variant of this adapter, when not being prompted by a reference image, can serve as an efficient control variate to reduce variance in score estimates, thereby enhancing both output quality and optimization stability. Our experiments demonstrate that the ISD loss consistently achieves visually coherent, high-quality outputs and improves optimization speed compared to prior text-to-3D methods, as demonstrated through both qualitative and quantitative evaluations on the T3Bench benchmark suite.
title ModeDreamer: Mode Guiding Score Distillation for Text-to-3D Generation using Reference Image Prompts
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.18135