neuroimage_denoiser - Neuroimage Denoiser

This is the best performing model, trained on iGluSnFR3 data (acquired at 100Hz). Settings used for training this model:

• cropsize: 32x32px
• noise center: 0
• noise scale: 2.0
• loss function: smooth L1
• y_train smoothed with Gaussian: yes
• Gaussian sigma: 1.0

- Python
Published by s-weissbach about 2 years ago

neuroimage_denoiser - model_weights_iglusnfr3

U-Net Model Weights These weights correspond to the trained U-Net model to denoise iGlu-Snfr3 data.

- Python
Published by s-weissbach over 2 years ago