I have a dataset of many 2-D line plots that represent height versus position measured from an AFM.

I am interested in training a model by classifying these lines into two groups: presence of a defect and no presence.

I believe a neural network would be appropriate for this since there are other features in the line plots that I want to be able to ignore.

The problem is that I am just starting by taking a certification course on machine learning but that hasn't given me enough information to develop one on my own. I have worked with a segnet based neural network but at a very user oriented level.

So I was wondering if anyone knew of an open-source algorithm that I could feed these lines into (maybe as vectors?) to train a model?

  • $\begingroup$ What is AFM short for? $\endgroup$ – David Cian Feb 16 at 23:26
  • $\begingroup$ Would you call this an image classification problem? $\endgroup$ – Dave Feb 17 at 0:27
  • $\begingroup$ AFM stands for atomic force microscopy, basically it just gives me height measurements of a sample along a straight line. $\endgroup$ – James Delles Feb 17 at 2:40
  • $\begingroup$ I wouldn't say it is an image classification problem because I was hoping to just use vectors holding the data for training a model, I don't know if training it as an image classification would be correct since the vast majority of the image is essentially blank which will heavily skew the training. $\endgroup$ – James Delles Feb 17 at 2:42
  • $\begingroup$ What is a defect in this task, is it about the position/orientation of the line or about how noisy the lines are? Your dataset is annotated, right? (I mean you have the correct label for every image) $\endgroup$ – Erwan Feb 17 at 16:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.