Question about transforms.Invertd (roi mask in original image)

[XIANYUNYEHE-DEL]

Hi, if i use transforms.CropForegroundd like this

transforms.CropForegroundd(keys=["image", "label"], source_key="label"),

label stands for mask.(For example, using mask to get a smaller region in original image for segmentation)
How do I set the parameters in the transforms.Invertd in inference stage.

[wyli]

there are some examples for the transform inverse :

• https://github.com/Project-MONAI/tutorials/blob/main/3d_segmentation/torch/unet_inference_dict.py
• https://github.com/Project-MONAI/tutorials/blob/main/modules/inverse_transforms_and_test_time_augmentations.ipynb