How to make sign symbols after hand detection through haar classifier hand?
After spending a couple of days i learnt the environment of opencv. i am going to create an android application for deaf and dumb in which i need to recognize sign symbols and convert into text. i am using haar classifier hand and successfully recognized hand further i am confused what to do next through hand classifier how to make sign symbols.. Any experts suggestion would be very helpful. Thanks in advance.
umm, does your cascade detect a thumbs-up gesture ? (or any other pose of your hand ?)
(using cascades for this might not be the best idea there is)
No it is only detect hand
so, you would need to train 1 cascade per gesture ? and then sort out the results ? seems a bad idea.
openpose might be a better idea
Thank you for your reply dear. can you please elaborate what is openpose? if i use tensorflow image classification? will it be beneficial? i need to recognize every sign and convert it to text or speech.
Yasir, there'll be a couple of problems in your way:
you NEED to do some research on your own here. for openbr, see the link above.
and again, Cascade-Classifiers will not solve your problem, sad as it is.
Yes i am following ASL (letter based) if i capture every letter sign i classify through tensor flow will it helpful? I think i should move to tensor flow from opencv it might be helpful. thank you for your time if you know something about tensor flow please suggest me how can i resolve my problem.
unfortunately, i'm not an expert in tf, or similar frameworks.
sure, you could train a multi-class classifier with it (and it's already better than the Cascade idea), but you'd still need a ton of labelled data for this.