Benchmarking Counterfactual Image Generation

TL;DR

This paper introduces a comprehensive benchmarking framework for counterfactual image generation, comparing various models across datasets and causal graphs, and providing a user-friendly Python package for future research.

Contribution

It presents a unified benchmarking framework and extends existing models to new datasets and causal graphs, highlighting Hierarchical VAEs as particularly effective.

Findings

Hierarchical VAEs outperform other models on most datasets and metrics.

The framework enables fair comparison of counterfactual image generation methods.

The Python package facilitates community extension and evaluation.

Abstract

Generative AI has revolutionised visual content editing, empowering users to effortlessly modify images and videos. However, not all edits are equal. To perform realistic edits in domains such as natural image or medical imaging, modifications must respect causal relationships inherent to the data generation process. Such image editing falls into the counterfactual image generation regime. Evaluating counterfactual image generation is substantially complex: not only it lacks observable ground truths, but also requires adherence to causal constraints. Although several counterfactual image generation methods and evaluation metrics exist, a comprehensive comparison within a unified setting is lacking. We present a comparison framework to thoroughly benchmark counterfactual image generation methods. We integrate all models that have been used for the task at hand and expand them to novel datasets and causal graphs, demonstrating the superiority of Hierarchical VAEs across most datasets and metrics. Our framework is implemented in a user-friendly Python package that can be extended to incorporate additional SCMs, causal methods, generative models, and datasets for the community to build on. Code: https://github.com/gulnazaki/counterfactual-benchmark.