search

qc17-THU / DL-SR

Tensorflow/keras implementation for image transformation from low-resolution (LR) image to super-resolved one, including single wide-field (WF) image super-resolution prediction and SIM reconstruction.
MIT License
98 stars 28 forks source link

Question about the dataset #8

Closed Luchixiang closed 2 years ago

Luchixiang commented 2 years ago

Hi. Thank you for your great work. I notice that just like what you said in the paper. Each set of raw SIM images is averaged to get the WF image while reconstructing to get the SR-SIM. I'm wondering in the BioSR dataset, which one is the reconstructed image(RawSIMData_gt.mrc or SIM_gt.mrc)? And which one did you use as the gtto train the network? Thank you Jasper

image
Luchixiang commented 2 years ago

Hi! Sorry to bother you again. I still have one question about data preprocessing. I find that there are two SIM_GT files in the F-action-nonlinear folder i.e. SIM_gt_a.mrc and SIM_gt_b.mrc. I'm wondering which one did you use to train the network?

qc17-THU commented 2 years ago

Hi! SIM_gt.mrc is the final SIM reconstructed SR image, which is used for training, and the RawSIMData_gt.mrc is the raw SIM images of the same ROI. All CNN models in our Nature Methods manuscript were training in an end-to-end manner aiming at directly outputing SR images. We use SIM_gt_a.mrc to train the network, while SIM_gt_b.mrc is provided for an overview of all ROIs with a gamma value of 0.6

Luchixiang commented 2 years ago

Thanks!