Model Learning in Image Analysis

[davpru96]

Hello everyone,
How can I apply a model learning after a segmentation in the image analysis?
I am trying to build a workflow which is able to count and frame cars in still imagery, and I used an INTERACTIVE ANNOTATOR node and an IMAGE SEGMENT FEATURES node in a LOOP to complete the feature extraction of all images.

I tried to use the simple regression tree learner and simple regression tree predictor, but I think that they do not perform correctly with images.
Thank you in advance!

[stelfrich]

Hi @davpru96,

@davpru96

I tried to use the simple regression tree learner and simple regression tree predictor, but I think that they do not perform correctly with images.

The simple regression tree learner does not do pixel-based classification, which is what you are looking for. You can achieve this with Deep Learning

Semantic Segmentation with Deep Learning in KNIME – knime

or by training a classifier on pixel features: https://www.knime.com/supervised-image-segmentation.

We are hosting a workshop about Deep Learning for Image Analysis next week, May 12: Feel free to attend!

Best regards,
Stefan

[davpru96]

Thank you!!!
I tried to run the second workflow to understand how it works, but I got this error. How can I solve it?

[davpru96]

I have already tried to follow the guidelines reported in the “Semantic Segmentation with Deep Learning” using the Classification Models on Boundaries. Then I do not know how to train the workflow, should I add the second part (from image preprocessing node) of this workflow (https://www.knime.com/supervised-image-segmentation) to mine?

Sorry to ask you so many questions, but I am new in KNIME and I am trying to figure out how it works.
Thanks again