Imagination Engine II

My long-term motivation is enabling humans to share what we imagine. My group and I (the “BCI team”) explored various methods and algorithms of image reconstruction, and eventually settled on electroencephalography (EEG).

At the Augmentation Residency, I explored the topic of image reconstruction from non-invasive brain recordings. This led us to a collaboration with Nataliya Kosmy’na from the MIT Media Lab, and her student Sven Günther. We collected over 20h of EEG data from ourselves with an 8-channel EEG headset. More on their research can be found here: Image Classification and Reconstruction from Low-Density EEG.  

My friend and co-resident later used the DreamDiffusion model that we explored to develop Imagination Engine I, an EEG-based application to generate art from brain waves using generative AI.