Heat Diffusion Models -- Interpixel Attention Mechanism
This work addresses image generation quality for computer vision applications, representing an incremental improvement over existing diffusion models.
The paper tackles the problem of preserving image details in diffusion models by proposing the Heat Diffusion Model (HDM), which incorporates an interpixel attention mechanism based on the heat equation, resulting in higher-quality image generation compared to models like DDPM, CDM, LDM, and VQGAN.
Denoising Diffusion Probabilistic Models (DDPM) process images as a whole. Since adjacent pixels are highly likely to belong to the same object, we propose the Heat Diffusion Model (HDM) to further preserve image details and generate more realistic images. HDM essentially is a DDPM that incorporates an attention mechanism between pixels. In HDM, the discrete form of the two-dimensional heat equation is integrated into the diffusion and generation formulas of DDPM, enabling the model to compute relationships between neighboring pixels during image processing. Our experiments demonstrate that HDM can generate higher-quality samples compared to models such as DDPM, Consistency Diffusion Models (CDM), Latent Diffusion Models (LDM), and Vector Quantized Generative Adversarial Networks (VQGAN).