I am new to datasets. I have got an annotations train.json for MR image data like this -

annonations json

I want to train a Yolo-V8 model using this MR data images extracted from dicom raw data and annotations for detection and localization. But I am not able to make any head from this bbox description. I am used to seeing "bbox":[x,y,width,height] type of description. The link to the dataset is THIS. I have uploaded the complete test.json, train.json, val.json here. please help.


1 Answer 1


The format description can be found on file DATASET.md at the github repo you linked:

The bounding box (bbox) format is [top left X position, top left Y position, top left Z position, deltaX, deltaY, deltaZ]. NOTE: X corresponds to row (SI), Y to column (AP), and Z (RL/LR) to depth.

  • $\begingroup$ Thank you so much and apologies, I should have read the documentation properly $\endgroup$
    – mrin9san
    Apr 1 at 10:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.