Conditional Balance: Improving Multi-Conditioning Trade-Offs in Image Generation

TL;DR

This paper introduces a method to improve the balance between content fidelity and artistic style in image generation by controlling attention layers in DDPMs, leading to better stylization quality.

Contribution

It identifies sensitive attention layers in DDPMs and directs conditional inputs there, enabling fine-grained control over style and content balance.

Findings

Enhanced stylization quality in generated images

Better alignment of style and content

Reduced issues from over-constrained inputs

Abstract

Balancing content fidelity and artistic style is a pivotal challenge in image generation. While traditional style transfer methods and modern Denoising Diffusion Probabilistic Models (DDPMs) strive to achieve this balance, they often struggle to do so without sacrificing either style, content, or sometimes both. This work addresses this challenge by analyzing the ability of DDPMs to maintain content and style equilibrium. We introduce a novel method to identify sensitivities within the DDPM attention layers, identifying specific layers that correspond to different stylistic aspects. By directing conditional inputs only to these sensitive layers, our approach enables fine-grained control over style and content, significantly reducing issues arising from over-constrained inputs. Our findings demonstrate that this method enhances recent stylization techniques by better aligning style and content, ultimately improving the quality of generated visual content.