Color Alignment in Diffusion
This work addresses the problem of precise color control in image synthesis for applications requiring specific color patterns, representing an incremental improvement over existing methods.
The paper tackles the challenge of fine-grained color conditioning in diffusion models by introducing a color alignment algorithm that projects diffusion terms into a conditional color space to align with input color patterns, achieving state-of-the-art performance in color pixel control while maintaining generation quality and diversity comparable to regular diffusion models.
Diffusion models have shown great promise in synthesizing visually appealing images. However, it remains challenging to condition the synthesis at a fine-grained level, for instance, synthesizing image pixels following some generic color pattern. Existing image synthesis methods often produce contents that fall outside the desired pixel conditions. To address this, we introduce a novel color alignment algorithm that confines the generative process in diffusion models within a given color pattern. Specifically, we project diffusion terms, either imagery samples or latent representations, into a conditional color space to align with the input color distribution. This strategy simplifies the prediction in diffusion models within a color manifold while still allowing plausible structures in generated contents, thus enabling the generation of diverse contents that comply with the target color pattern. Experimental results demonstrate our state-of-the-art performance in conditioning and controlling of color pixels, while maintaining on-par generation quality and diversity in comparison with regular diffusion models.