An Gesture Recognition Project using OpenCV 3.3.0, Numpy 1.13.3 and Python 3.6.3 for detecting and marking pre-defined hand gestures using various image processing techniques such as Background Subtraction, Colorspace Conversion, Gaussian and Median Blur Smoothening, Histogram Projection, Thresholding, Contour and Convex Hull Finding and more!
For a one stop solution, I always use conda. First install conda and create you own environment as:
conda create --name myenv
where myenv is the environment name (which can be changed at will). Then activate the environment by:
source activate myenv
and then install OpenCV and Python using a single command:
conda install -c conda-forge opencv
You can execute the project using the following command in the src folder:
python3 main.py
Currently, the project is in a standalone working phase (but without a number of possible optimisations). This project is the culmination of more than 6 months of brainstorming (read as 20 core days) through various OpenCV tutorials and the library documentation, auditing freely available lectures on image processing (thanks to Prof. Ranga Rodrigo of University of Moratuwa, Sri Lanka), losing and rediscovering motivation time and again and making short progresses, and looking up at different codebases for tackling some minor problems.
The project currently works in HSV(Hue-Saturation-Value) colorspace for optimal results under thresholding and background subtraction.
Any suggestions and PRs are always welcome!
(Please ignore the awkward camera angles. It is a well documented camera placement error on the XPS 15 9560.)
