search

RomGai / BrainVis

Official code repository for the paper: "BrainVis: Exploring the Bridge between Brain and Visual Signals via Image Reconstruction"
https://brainvis-projectpage.github.io/
MIT License
44 stars 2 forks source link

Question on SSIM calculation #10

Closed minsuk00 closed 3 months ago

minsuk00 commented 3 months ago

Thank you for the nice paper. Could you provide details on how you calculated the SSIM scores? 0.75 seems incredibly high for EEG-based image generation, which I'm having difficulties reproducing. Most fMRI-based methodologies produce around 0.35 and other EEG based papers have around 0.2~0.3. Thank you!

minsuk00 commented 3 months ago

I see! Thank you for confirming.

RomGai commented 3 months ago

We have found that our script for calculating SSIM has something wrong. It may probably because the pre-processing (for high-level evaluation) parts in our code. Considering we do not have any low-level design, we delete this script and replaced SSIM by FID. The paper will also be updated later.