Benchmarking Counterfactual Image Generation
This work addresses the need for standardized evaluation in counterfactual image generation, which is crucial for realistic edits in domains like natural or medical imaging, though it is incremental as it builds on existing methods and metrics.
The authors tackled the problem of evaluating counterfactual image generation methods, which lack observable ground truths and require adherence to causal constraints, by developing a comprehensive benchmarking framework. Their results demonstrated the superiority of Hierarchical VAEs across most datasets and metrics, with the framework implemented in a user-friendly Python package for community use.
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.