The Ebbinghaus Benchmark is an open-source project I've put together to allow researchers a simple yet flexible API to generate variants of the Ebbinghaus Illusion. The full project is on GitHub and Kaggle , and is comprised of two additional phases: Evaluation and Analysis. It is suitable both as an AI benchmark and for human participants.

You can also find a Python implementation and a notebook example with open-source, free models.

In this application I provide you with a UI to generate the underlying table as well as the stimuli. Please read more about the different options in my repository and contact me if you have further questions.

I hope you find it useful.

Yann.
yann-dev.io