Uni-ControlNet: All-in-One Control to Text-to-Image Diffusion Models
Shihao Zhao† The University of Hong Kong、Dongdong Chen Microsoft、Yen-Chun Chen Microsoft、Jianmin Bao Microsoft、Shaozhe Hao The University of Hong Kong、Lu Yuan Microsoft、Kwan-Yee K. Wong∗ The University of Hong Kong Corresponding Author, ††\dagger Intern at Microsoft
Abstract
Text-to-Image diffusion models have made tremendous progress over the past two years, enabling the generation of highly realistic images based on open-domain text descriptions. However, despite their success, text descriptions often struggle to adequately convey detailed controls, even when composed of long and complex texts. Moreover, recent studies have also shown that these models face challenges in understanding such complex texts and generating the corresponding images. Therefore, there is a growing need to enable more control modes beyond text description. In this paper, we introduce Uni-ControlNet, a unified framework that allows for the simultaneous utilization of different local controls (e.g., edge maps, depth map, segmentation masks) and global controls (e.g., CLIP image embeddings) in a flexible and composable manner within one single model. Unlike existing methods, Uni-ControlNet only requires the fine-tuning of two additional adapters upon frozen pre-trained text-to-image diffusion models, eliminating the huge cost of training from scratch. Moreover, thanks to some dedicated adapter designs, Uni-ControlNet only necessitates a constant number (i.e., 2) of adapters, reg
中文速览
文本生成图像（Text-to-Image）扩散模型虽然效果出色，但单靠文字描述很难精准控制图像的细节布局，因此研究者们开始探索如何把边缘图、深度图、分割掩码等额外信号引入生成过程。现有方法（如ControlNet）的问题在于每增加一种控制条件就要新增一个独立适配器，导致训练成本和模型体积随条件数量线性膨胀，多条件组合使用时效果也不理想。本文提出Uni-ControlNet，将各类控制信号归为"局部控制"和"全局控制"两大类，只需在冻结的预训练扩散模型基础上额外训练两个专门设计的适配器——局部适配器采用多尺度特征归一化注入策略，全局适配器则将参考图像的CLIP嵌入拼接到文本token中通过交叉注意力发挥作用——无论使用多少种条件，适配器数量始终保持为2个。实验表明，Uni-ControlNet在可控性、生成质量和多条件组合能力上均优于ControlNet等现有方法，而所需训练数据仅1000万图文对、训练1个epoch，大幅降低了研究门槛，对实际部署更为友好。
原文 arXiv:2305.16322;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2305.16322v3