Saved in:
Bibliographic Details
Main Authors: Wang, Junyan, Sun, Zhenhong, Tan, Zhiyu, Chen, Xuanbai, Chen, Weihua, Li, Hao, Zhang, Cheng, Song, Yang
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2403.05239
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929268661944320
author Wang, Junyan
Sun, Zhenhong
Tan, Zhiyu
Chen, Xuanbai
Chen, Weihua
Li, Hao
Zhang, Cheng
Song, Yang
author_facet Wang, Junyan
Sun, Zhenhong
Tan, Zhiyu
Chen, Xuanbai
Chen, Weihua
Li, Hao
Zhang, Cheng
Song, Yang
contents Vanilla text-to-image diffusion models struggle with generating accurate human images, commonly resulting in imperfect anatomies such as unnatural postures or disproportionate limbs.Existing methods address this issue mostly by fine-tuning the model with extra images or adding additional controls -- human-centric priors such as pose or depth maps -- during the image generation phase. This paper explores the integration of these human-centric priors directly into the model fine-tuning stage, essentially eliminating the need for extra conditions at the inference stage. We realize this idea by proposing a human-centric alignment loss to strengthen human-related information from the textual prompts within the cross-attention maps. To ensure semantic detail richness and human structural accuracy during fine-tuning, we introduce scale-aware and step-wise constraints within the diffusion process, according to an in-depth analysis of the cross-attention layer. Extensive experiments show that our method largely improves over state-of-the-art text-to-image models to synthesize high-quality human images based on user-written prompts. Project page: \url{https://hcplayercvpr2024.github.io}.
format Preprint
id arxiv_https___arxiv_org_abs_2403_05239
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Towards Effective Usage of Human-Centric Priors in Diffusion Models for Text-based Human Image Generation
Wang, Junyan
Sun, Zhenhong
Tan, Zhiyu
Chen, Xuanbai
Chen, Weihua
Li, Hao
Zhang, Cheng
Song, Yang
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Vanilla text-to-image diffusion models struggle with generating accurate human images, commonly resulting in imperfect anatomies such as unnatural postures or disproportionate limbs.Existing methods address this issue mostly by fine-tuning the model with extra images or adding additional controls -- human-centric priors such as pose or depth maps -- during the image generation phase. This paper explores the integration of these human-centric priors directly into the model fine-tuning stage, essentially eliminating the need for extra conditions at the inference stage. We realize this idea by proposing a human-centric alignment loss to strengthen human-related information from the textual prompts within the cross-attention maps. To ensure semantic detail richness and human structural accuracy during fine-tuning, we introduce scale-aware and step-wise constraints within the diffusion process, according to an in-depth analysis of the cross-attention layer. Extensive experiments show that our method largely improves over state-of-the-art text-to-image models to synthesize high-quality human images based on user-written prompts. Project page: \url{https://hcplayercvpr2024.github.io}.
title Towards Effective Usage of Human-Centric Priors in Diffusion Models for Text-based Human Image Generation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2403.05239