How to import DICOM images in Knime Deep learning flowchart and to merge them with excel data file.

KNIME Extensions
1

Greetings,
I am trying to import DICOM images or nifti images in deep learning.
Do I require another extension or another format? Is there a node which can help me with that?
Also I want to know how to link these DICOM images to the excel datasheet.

With regards
Akshaykumar

2

Hi @Akshaykumar,
to use our Deeplearning extension with image data, you need to additionally install the KNIME Image Processing - Deep Learning Extension.

You should be able to read both DICOM and nifti images using the Image Reader node. You can read data from a excel sheet using the Excel Reader node. Once the data is in KNIME you can join the images with the data from the excel sheet, e.g. using a Joiner node.
best,
Gabriel

1 Like
3

@gab1one Thank you very much.
I did what you suggested and came it the point of running the deep learning model.
But it did not run,
Showing the error “Length of current input does not match expected length. Possible images or collections may not be of same size.”
I think I need to figure out how to match image sizes of dicom data, I was thinking maybe it would be uniform.
Still I am very happy that I reached till the final step.

Cheers,
Akshaykumar

4

Hello Akshaykumar,

the error indicates that your images have varying sizes, but we require images of the same size to build our batches for training and execution.
There are different strategies to tackle this problem:

  • Use the Image Resizer to enforce a common size of your images
  • Use the Tiling Loop to extract patches of the same size from all your images

As the appropriate approach depends to a large degree on your use-case, I’d have to know more about what you want to achieve in order to provide further guidance.

Cheers,
nemad

1 Like
5

Thanks I used Image resizer but I think Tiling loop would be better,
I am trying to classify MRI image data. But problem is some patients have tilted their heads in scan making it difficult to resize them all.
I could not run deep learning node because of this error “ERROR DL4J Feedforward Learner (Classification) 0:14 Execute failed: Not enough memory available for DL4J. Please consider increasing the ‘Off Heap Memory Limit’ in the DL4J Prefernce Page.”
Please help.

With regards
Akshaykumar

6

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.