Show HN: ASL Classifier Built with CoreML and Roboflow
1–10 of 17 posts
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#2Neat! I would qualify that this is an ASL fingerspelling classifier, rather than a general classifier for the ASL language.
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#3Did you find the model to perform better / worse when the background varied? I see it's all wood table examples in the gifs.
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#4Is the app available on the App Store?
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#5pretty awesome - I wonder if the algo could be used to teach people how to sign the alphabet - new twist on machine "learning"
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#6pretty awesome - I wonder if the algo could be used to teach people how to sign the alphabet - new twist on machine "learning"
You could probably do that!
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#7Did you find the model to perform better / worse when the background varied? I see it's all wood table examples in the gifs.
Yeah; I found that if there was a lot of other objects in view (laptop/phone/notebook/pens/etc) then it had a hard time (I'm not great at signing anyways, ha
It may have performed better if I had done some image segmentation processing before running inference, then it may have performed even better
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#8Is the app available on the App Store?
Not at the moment! Just an open-source demo for now; but I think this could be a very cool app for whenever the Apple glasses get released...
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#9Very cool project.
I love these applications of CV for accessibility for the blind. This is exactly what technology is made for.
Re: Show HN: ASL Classifier Built with CoreML and Roboflow
#10Very cool project. I love these applications of CV for accessibility for the blind. This is exactly what technology is made for.
Thanks Lenny! And yes, agreed. You should check out Oz Ramos' work: https://handsfree.js.org/