How to configure label for image classification

[matsujirushi]

Question/Issue:
I believe that to perform image classification, I need to upload images and assign labels to them. However, I’m not sure how to add or configure the labels.

I set the labels when capturing images with the smartphone app, but it seems that these labels are not being recognized as labels during transfer learning.

According to this tutorial, there should be a “LABEL” column, but in my dataset I only see a “LABELS” column, which appears to be intended for object detection instead.

Image classification - Edge Impulse Documentation

How to configure label for image classification?

Project ID:

Context/Use case:

Steps Taken:

Expected Outcome:

Actual Outcome:

Reproducibility:

• [x] Always
• [ ] Sometimes
• [ ] Rarely

Environment:

Logs/Attachments:

Additional Information:

[matsujirushi]

[marcpous]

Hello @matsujirushi you need to go to the tab Labeling queue. There you have 60 images to label.

[matsujirushi]

Hi @marcpous ,
Thanks for reply.

When I switch to the “Labeling queue” it takes me to a screen where I can add bounding boxes to images.
I want to perform IMAGE CLASSIFICATION rather than OBJECT DETECTION.
In that case, do I still need to set bounding boxes?

[matsujirushi]

There was a “labeling method” setting in the dashboard. After switching it to “one label per data item”, it started behaving exactly as I expected.

[marcpous]

Exactly! This makes the trick @matsujirushi

Thanks for sharing the solution here

Let us know if we can help you more!