HIGHLIGHTS
- who: Lynn Le from the Northwest University, China have published the Article: Brain2Pix: Fully convolutional naturalistic video frame reconstruction from brain activity, in the Journal: (JOURNAL)
- what: The authors wanted to focus on early visual areas, so the authors trained the model on V1, V2, and V3 individually (which the authors called the ROI experiment) using the retinotopic maps provided by the paper that provided the dataset (Seeliger et_al, 2019). The poor performance of the model based on V1 can be due to two main reasons: Either the adversarial and feature losses have larger weights . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.