Final screenshots of part 2 ,And end of tutorial
Output added to captured image(Another window)
So the above hand "gesture" I have tried to capture is a very rough one ,not at all all right with its orientation and its not correctly printing defects all round.
The problem is its really troublesome to set up proper background and many related things plus there are some issues with cameras which seems to change based on the camera hardware
while solution exist to them and are not that difficult to implement in terms of days but honestly I dont see any point in continuing further. Honestly ,gesture recognition and python are overrated in my opinion. Atleast from what I know the opencv libary is open source,so while it being open source has its benefits ,its drawbacks are a lot, it doesn't have the driving force that often comes with a corporate i.e to get results. so that means there are many ,you could say "rough edges" with it ,you cannot customize the classes in it using python at all. Its coded entirely in c++. It doesn't have a wrapper which may have been used to add some customizations... I suppose that would be limited but ...would solve my functions.
I suppose as the reputation goes python is useful for application which benefit from fast results.
No wonder,scientific computing and data science etc finds python very useful. Honestly its like they are made for each other....(Okay that might be a little farfetched). But the point is as long as you don't care about robustness and daily use of the same program and you probably want to focus only the scientific experiments then they might be useful...but even then it still wouldn't compare to corporate products...obviously I guess..
One of the benefits I found from python ,is that you can create your own "experimental setup" in your computer without depending on corporate products. Its hardly takes time(like in terms of a few weeks at worst)Its because python so many libraries to manipulate the software and hardware connected to your laptop or pc. Python's libraries have been around for many years. so that means it has a lot of time to mature . Perhaps that may be why python was preferred choice for the AI and scientific community.
You could probably use it for serious experiments too . I am sure there must be libraries to handle collaboration over a network in python ,so it must not be that hard to use the wrapper API to manipulate file transfers over a network and other things (like encryption etc).
I honestly felt gesture recognition was overrated. From my experience with openCV ,while the few algorithms i had some experience with ,were definitely high quality ( I am using that word because it was based on a research paper) But the point i am making is ,that in my college or in most colleges in general, its a very popular topic and there is often a rush or a "enthusiasm" towards implementing or making it, even if its not related with the daily computer applications they will most probably be dealing with ...... I mean maybe in some other part of the world , I bet all of this must be very useful, but in general (software and real) life ,I hardly see any point in it.
I am not talking about it from the context of my paper ,but just in general. Same goes for machine learning and data science.
I am thinking of few interesting topics like android and ios and I want to explore a few more things related to desktop, of handling different outputs from different commonly used software(by me) and manipulating other forms of output(like mouse cursor ,sound) etc so also other forms of hardware like
(USB , touchpad ,mouse,speakers,microphone) I don't feel like touching internal hardware... I am just going to leave that to the other experts .... 😆
I developed the code very quickly ....I am not talking about the entire project, but basically that because of the internet and the maturity of python with so many opinions around for each topic and a relatively good documentation ,its fantastic tool for prototyping. Really for daily small things I think python can be inmmensly useful.
I will also upload my code on github.( I should also revise this post ....so many grammer mistakes)
uploaded
Note:For understanding the code,there are many comments in the code,see if you are able to understand using that
Comments
Post a Comment