This question has been asked before here but never received an answer.

I'm planning to write a bunch of components that do some processing on single images. However, I have many such images and would like to find a way to iteratively run over a list of such images and execute that single-image component.

Is this at all possible in Orange3?

If not, do you have any ideas about how to build a component for that? I was thinking, perhaps it's possible to create a component "Repeat" that takes a single input (e.g., a list of image file paths) and repeatedly generates an output, or refreshes its single output repeatedly with new data. Or something along those lines.

Alternatively, I have to create a higher-level component that just takes a list of image file paths, processes them and outputs a new list of file paths pointing to the processed images. But from a OO-perspective, this component does too much and will be getting too complex. I'd rather keep the components simple and "atomic".

Does this make sense?

  • Do you want to use Orange as a software or write your own Python script that uses the Orange3 library? Can you tell what analysis do you want to do with Orange? – Primoz Oct 8 at 10:54
  • Thanks for your reply. I want to use Orange as a visual editor for building processing pipelines. My analysis is not really of concern here because this is related to preprocessing, which comes before analysis. For example, filtering an image where you set each non-zero pixel to 1 and leave the others to zero thereby creating a binary mask. Such an operation I would like to perform repeatedly on many images, not just one. I hope that makes a bit more sense. – Ralph Oct 9 at 3:49
  • If you're interested in general about the analyses I do, its radiomics on medical images :) – Ralph Oct 9 at 3:50
  • Ok, I understand now. Orange has a small set of widgets to work with images. They are included in the image analytics add-on which can be installed from the Options>Add-ons menu. Currently, Orange support loading the images, computing embedding (vector representation of the content), viewing the images and grouping them in the grid. If you want to have custom widgets for your functionalities (filtering, ...) you are welcome to implement them and submit as a pull request on GitHub github.com/biolab/orange3-imageanalytics. – Primoz Oct 9 at 11:48
  • Thnx, I will think about it :) – Ralph Oct 9 at 11:50

Your Answer

 

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.