MIC-DKFZ / medicaldetectiontoolkit

The Medical Detection Toolkit contains 2D + 3D implementations of prevalent object detectors such as Mask R-CNN, Retina Net, Retina U-Net, as well as a training and inference framework focused on dealing with medical images.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

When using batch_generator, where should I put "boxes_coords" in?

fmo-voxelcloud opened this issue · comments

I have a custom datasets in 3D, labeled with segmentation and I did ConvertSegToBoundingBoxCoordinates already, so I have a annotation file (.json), contains each image's path, segmentation path and b-boxes.

Now I am confusing when re-writing preprocessing.py script, as showed in README, I should have a dict contains:
"data": preprocessed image (.npy format)
"seg": preprocessed segmentation image (.npy format)
"pid": patient id
"class_target": a list of each ROI's class, for example, [1, 2] (means this image have 2 rois, first is class_1 and second is class_2)

So, where the bounding boxes coordinates should be put ??

Hello! Did you manage to write a customized preprocessing.py for your own dataset? If yes, could you please provide it? Thanks in advance.