A tutorial- Gesture recogniton (Thoughts and some plans and steps) Part 1
So I guess I am trying to explore things as of now .
So what I would say about this tutorial is that its more of an exploration endeavor and more of a learning opportunity for me ,to solidify some knowledge(or paradigms) I wanted to glean from the gesture recognition ,python and this project as a whole.
In the process, if this "tutorial" ends up being useful for someone else, then I guess its a win-win.
So I first went through this website
This website gives a vague(but a good) tutorial. This tutorial gives links and references and gives some brief explanations on some keywords.Along with that the code is also provided. The code given can be implemented in python 3 or 2 with OpenCV library. The OpenCV documentation was also very useful to me along with one more distribution known as the anaconda. But I suppose anaconda is a overkill for this.
But if you happen to get these tools then you can set up the implementation fairly quickly probably in 3-4 days. You will have to play around with the anaconda interface and may also need to install opencv library using pip command on the terminal. The Spyder IDE was very useful for me and I loved it .
The debug functions provided in the OpenCV library were very useful. So were all the transformation functions (image).
OpenCV also has functions to open webcam input. I haven't tested anything else.... like maybe having input from video call or maybe input from mobile. I suppose there may be many more ways.
But thats not what I wanted to focus on with this "tutorial".
What I wanted to focus on was refining the control over many of the features in the program as there are many features that give very useful output but are hidden otherwise. I want to control these outputs individually.
Some of these outputs are:
-Binary image
-Grayscale image
-Control the "modifications" on the original image ,like the convex hull, convexity defects .
-The contour drawing too.
Comments
Post a Comment