search

BeileiCui / EndoDAC

[MICCAI'2024] EndoDAC: Efficient Adapting Foundation Model for Self-Supervised Depth Estimation from Any Endoscopic Camera
30 stars 2 forks source link

obtain depth maps #11

Open Beilinlin opened 4 weeks ago

Beilinlin commented 4 weeks ago

Hi! How can I use it to obtain depth maps for a set of endoscopic images?

BeileiCui commented 3 weeks ago

You can either train the model follow the readme or use the checkpoint provided and use test_simple.py or evaluate_depth.py to obtain the depth map. You may need to modify the code to fit your data(test_simple.py can be used directly)

Beilinlin commented 2 weeks ago

Thanks！How should I use my own data for training?

BeileiCui commented 1 week ago

Thanks！How should I use my own data for training?

@Beilinlin The only thing you need to prepare is the video (or processed video frames) of your own data. Then I suggest you imitate the structure of the datasets/scared_dataset.py to construct your own dataloader and replace it in trainer_end_to_end.py to start your training.