Closed puallee closed 1 month ago
Hi,
Thanks for your interest in our work.
In our provided data, folder 0 comprises all the training data given in the dataset. We then generate various augmentations on the training data in folder 0 and store the augmented data in folders labelled from 1 to 40, each containing an augmented copy of the whole training set. You may refer to this comment for more information. To follow our procedure, you need to download the provided data and then follow the data preprocess instructions outlined here. This will then enable you to proceed with model training using our code.
Alternatively, you can compute SAM image embeddings online during training and conduct data augmentation online. This may require minor adjustments to the dataset code to switch to image input and incorporate specific augmentations.
can you provide the code for dataset augmentations
Hi,
Sorry for the late reply.
We achieved data augmentation in this line using the function augmentation
. The augmentation
function is defined here.
Please note that we achieve data augmentation in an offline manner. If you want to conduct data augmentation online, you will need to modify the code and add the augmentation part to the dataloader.
In your training stage, it need training sets 1,2,3 of endovis_2018 and endovis_2017,Can you update training set 1,2,3,?