search

ucwxb / I-MedSAM

[ECCV2024] I-MedSAM: Implicit Medical Image Segmentation with Segment Anything
24 stars 0 forks source link

Image size for training and inferencing #5

Open ljmkgbkq opened 2 hours ago

ljmkgbkq commented 2 hours ago

Hello. Are there any requirements for picture sizes for training and reasoning? I see image_size=384 in the code, but this size is a little small for my data set. Thank you

ucwxb commented 2 hours ago

Hi, the size of the pictures is not restricted. We follow the experiment settings of baseline SwIPE, which set 384 as the resolution of images. You can adjust the two params 'image_size' and 'label_size' in 'main.py' to fit your dataset.