Endo-FM
Endo-FM copied to clipboard
med-air
[MICCAI'23] Foundation Model for Endoscopy Video Analysis via Large-scale Self-supervised Pre-train
Can you provide the weights for all the comparative methods trained on your dataset? Thank you~
Hi, great work! But I have a question I don't understand. The backbone you used for training is a timesformer which takes a sequence of frames as input, but for...
Hi, I have a question about the normalization in the pretraining. Did you use standard Imagenet values to normalize or did you normalize according to the medical data domain? Thanks!!
This is great work! Are the other methods you compared also pretrained on the same union of 7 datasets as Endo-FM? Do all methods use the same set of hyperparameters?
Thanks for your great work! When I preprocess for SUN & SUN-SEG, I have a question for data/trans_videos_pretrain.py line 14-16: `datasets = { 'SUN-SEG': 0, }` only concludes the SUN-SEG,...
Hi, thanks for your nice work! I'm not sure how the fine-tuning is performed for CVC. As far as I understand from other issues, you feed the network with the...
I followed the guide in ReadMe and compile the STFT using a RTX4090. It successfully compiled but, when I run the finetuning, it outputs the following error: Traceback (most recent...
1. for pre-training FM, how many frames are used ? (150fr? for each batch)? 2. for downstream task, I see that you used 224 224 with 1 frame for segmentation....
Thanks for your work. I have a question about the metric used in TransUNet. In the /TransUNet/trainer.py file, I found the metrics defined as follows: def calculate_metric_percase(pred, gt): pred[pred >...
Hello, While reviewing your code for fine-tuning on a segmentation task, I had a couple of questions. I noticed that only train and test modes are defined in the dataset,...
