Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

transform torchio subject - mask with multiple masks and different shapes than image

See original GitHub issue

I have an RGB Image and a 7-class ground truth mask and I’m trying to figure out how to integrate torchio into my pipeline. I get this error

ValueError: Images in sample have inconsistent shapes:
{'Image': (1, 3, 4041, 2817), 'Mask': (1, 7, 4041, 2817)}

Also, when transforming the ground truth mask, what is the appropriate shape? onehot (eg. 1, 7, 4041, 2817) or categorical (eg. 1, 1 ,4041, 2817)?

Issue Analytics

State:
Created 3 years ago
Comments:11 (6 by maintainers)

Top GitHub Comments

1reaction

fepegarcommented, Jul 12, 2020

You can use image_interpolation='linear'.

Interpolation with B-splines of order 1 is equivalent to using linear interpolation, AFAIK. In 2 and 3D, linear interpolation is also called bilinear and trilinear interpolation, respectively.

0reactions

xen0f0ncommented, Jul 12, 2020

Yes. I though that Resample does resize… Although I resize 2d images, I think it works. I use the bspline interpolation method (since there is no bilinear or bicubic). Am I missing something?

Top Results From Across the Web

torchio.transforms.preprocessing.intensity.mask - Read the Docs

If ``None``, all values larger than zero will be used for the mask. ... image to mask >>> transformed = mask(subject) # Set...

TorchIO: A Python library for efficient loading ... - arXiv

We present TorchIO, an open-source Python library to enable effi- cient loading, preprocessing, augmentation and patch-based sampling of medical.

API Reference - ivadomed documentation

Compute a soft Dice coefficient between im1 and im2, ie equals twice the sum of the two masks product, divided by the sum...

TorchIO: a Python library for efficient loading ... - arXiv Vanity

In comparison, many medical images are reconstructed such that voxel location ... TorchIO provides several MRI-specific augmentation transforms related to ...

How to perform identical transform on both image and target ...

If you want to treat each mask separately as a sample, you could call view(-1, height, width) on it to put all masks...