Saved in:
Bibliographic Details
Main Authors: Xiao, Changming, Yang, Qi, Zhou, Feng, Zhang, Changshui
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2309.04109
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913524082540544
author Xiao, Changming
Yang, Qi
Zhou, Feng
Zhang, Changshui
author_facet Xiao, Changming
Yang, Qi
Zhou, Feng
Zhang, Changshui
contents Diffusion models have revolted the field of text-to-image generation recently. The unique way of fusing text and image information contributes to their remarkable capability of generating highly text-related images. From another perspective, these generative models imply clues about the precise correlation between words and pixels. In this work, a simple but effective method is proposed to utilize the attention mechanism in the denoising network of text-to-image diffusion models. Without re-training nor inference-time optimization, the semantic grounding of phrases can be attained directly. We evaluate our method on Pascal VOC 2012 and Microsoft COCO 2014 under weakly-supervised semantic segmentation setting and our method achieves superior performance to prior methods. In addition, the acquired word-pixel correlation is found to be generalizable for the learned text embedding of customized generation methods, requiring only a few modifications. To validate our discovery, we introduce a new practical task called "personalized referring image segmentation" with a new dataset. Experiments in various situations demonstrate the advantages of our method compared to strong baselines on this task. In summary, our work reveals a novel way to extract the rich multi-modal knowledge hidden in diffusion models for segmentation.
format Preprint
id arxiv_https___arxiv_org_abs_2309_04109
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle From Text to Mask: Localizing Entities Using the Attention of Text-to-Image Diffusion Models
Xiao, Changming
Yang, Qi
Zhou, Feng
Zhang, Changshui
Computer Vision and Pattern Recognition
Diffusion models have revolted the field of text-to-image generation recently. The unique way of fusing text and image information contributes to their remarkable capability of generating highly text-related images. From another perspective, these generative models imply clues about the precise correlation between words and pixels. In this work, a simple but effective method is proposed to utilize the attention mechanism in the denoising network of text-to-image diffusion models. Without re-training nor inference-time optimization, the semantic grounding of phrases can be attained directly. We evaluate our method on Pascal VOC 2012 and Microsoft COCO 2014 under weakly-supervised semantic segmentation setting and our method achieves superior performance to prior methods. In addition, the acquired word-pixel correlation is found to be generalizable for the learned text embedding of customized generation methods, requiring only a few modifications. To validate our discovery, we introduce a new practical task called "personalized referring image segmentation" with a new dataset. Experiments in various situations demonstrate the advantages of our method compared to strong baselines on this task. In summary, our work reveals a novel way to extract the rich multi-modal knowledge hidden in diffusion models for segmentation.
title From Text to Mask: Localizing Entities Using the Attention of Text-to-Image Diffusion Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2309.04109