i need help with an idea and i dont know where to start (i only know python)
I've had this idea for a app/website for a couple of months now the problem is i dont know where to start (and as i said in the title my knowledge is limited), the idea goes something like this: an app that can understand sign launguage and type it in the…
I've had this idea for a app/website for a couple of months now the problem is i dont know where to start (and as i said in the title my knowledge is limited), the idea goes something like this: an app that can understand sign launguage and type it in the screen as well as read it out loud automatically ,kinda like google lens where it can translate texts in real time but this app will use the camera and most likely a motion capturing system to translate sign launguage i dont know if it will be worthy or if im cappable of doing it but im willing to try my best (also sorry for any misspellings english isnt my first launguage) thanks in advance
Collected discussion
That's incredibly ambitious and probably not an ideal first project. Or 20th project.
I know but I want to try it
Thank you I will try
Oh thank you so much I didnt know! I will definitely check it out
You are so sweet!! Thank you
Guessing here, but opencv would probably be the initial thing to actually get the sign language bit.
This has already been done, but ofc try it yourself. If you Google it you should find a fairly details Medium post about it from 3 years ago.
Don't let the size of idea stop you , Every big project start with small steps.You would begin with python and openCV to recognise a few basic signs then you gradually add text output, speech and then a final a proper app. You will learn faster by building it step by step instead of making a final product from day one .
I’d probably point you towards building a machine learning model to identify the hand shapes. The rest is kinda a simple app or site to interact with the model. This stuff has easy enough tutorials out there, your biggest hurdle is a large and diverse enough training set to get meaningful results. That said, this might be a challenge, but I bet you could build a kinda shitty version of it.
This is very ambitious. 1.You need to decide web app Vs desktop. 2.Then language and framework that offer the best tools for the project. 3.What about different countries version's of sign language. 4.You need some type of ML programming to classify sign language gestures. That means making your own or using an existing one which you'll probably have to weigh how to get results (probably an API if one exists). 5.How to present the data back to user. Is it one full video analysis then paste translated or is it real time? Which means you may need to jump back to the past sentence and context fill/correct outputs. 6.Profanity filtering may be needed I'd probably go web app for cross platform use. But web development is its own beast. Have you researched tools to recognise sign language? Because if you have to make this yourself you'll need hundreds or thousands of hours for training data.