*|MC:SUBJECT|* *|MC_PREVIEW_TEXT|*

Hello!

With the big hype Deep Learning has been gaining in recent years, I found myself kind of *forced* to start employing it in my research. Well, I don't have any problem with that of course, on the contrary, it is indeed very interesting and have shown to solve many issues that traditional methods couldn't.

The dilemma I faced was that should I use Deep Learning in /image/ /segmentation/ or not? Especially that I have been utilizing image processing based techniques for such task with no issues so far. But again, everyone is talking about Deep Learning and it's kind of hard to imagine a publication without it at this point.


So, here I am, using Deep Learning in segmenting skin lesion images in my papers below:

*A Deep Learning Based Approach to Skin Lesion Border Extraction With a Novel Edge Detector in Dermoscopy Images* <https://dspace.stir.ac.uk/retrieve/21516e25-9ceb-41e5-a514-295a26c4120c/IJCNN%20paper.pdf>

*Supervised Versus Unsupervised Deep Learning Based Methods for Skin Lesion Segmentation in Dermoscopy Images* <https://dspace.stir.ac.uk/bitstream/1893/29692/1/CanadianAI19_submitted.pdf>

I cannot say that the results are bad. But, to be honest, I returned back to my old habit in subsequent papers (which I will share with you soon) in using image processing based techniques for skin lesion segmentation. And, guess what? They actually outperformed U-Net at least and I have made such comparison clear in the papers to highlight that.

Well, that's tt for now...

You can kindly find our other work related to skin lesions and melanoma at our project website: *DeepDerma* <https://www.deepderma.ai/>.

Oh, didn't I say that we are always open for collaboration. If you would like to work with us or have any thoughts or projects you would like me to get involved with you in, please don't hesitate to shoot me an email and we can take it from there.

Best,
Abder-Rahman
*abder.io* <https://abder.io/>


_______________________________________________
uai mailing list
[email protected]
https://secure.engr.oregonstate.edu/mailman/listinfo/uai

Reply via email to