I2AM: Interpreting Image-to-Image Latent Diffusion Models via Bi-Attribution Maps

TL;DR

This paper presents I2AM, a novel method for visualizing bidirectional attribution maps in image-to-image diffusion models, improving interpretability and aiding model debugging across various tasks.

Contribution

Introduces I2AM, a new approach for interpreting I2I diffusion models by visualizing cross-attention-based attribution maps and proposing the IMACS metric for evaluation.

Findings

I2AM effectively highlights key regions influencing output generation.

IMACS correlates strongly with existing performance metrics.

I2AM facilitates model debugging and refinement.

Abstract

Large-scale diffusion models have made significant advances in image generation, particularly through cross-attention mechanisms. While cross-attention has been well-studied in text-to-image tasks, their interpretability in image-to-image (I2I) diffusion models remains underexplored. This paper introduces Image-to-Image Attribution Maps (I2AM), a method that enhances the interpretability of I2I models by visualizing bidirectional attribution maps, from the reference image to the generated image and vice versa. I2AM aggregates cross-attention scores across time steps, attention heads, and layers, offering insights into how critical features are transferred between images. We demonstrate the effectiveness of I2AM across object detection, inpainting, and super-resolution tasks. Our results demonstrate that I2AM successfully identifies key regions responsible for generating the output, even in complex scenes. Additionally, we introduce the Inpainting Mask Attention Consistency Score (IMACS) as a novel evaluation metric to assess the alignment between attribution maps and inpainting masks, which correlates strongly with existing performance metrics. Through extensive experiments, we show that I2AM enables model debugging and refinement, providing practical tools for improving I2I model's performance and interpretability.