Given an image containing handwritten numbers, this package segments the image and identifies each digit present in the image. Shortly, a handwritten digit recognizing software(a very crude one though). The underlying neural network can be retrained to meet specific needs ##Documentation
To access the documentation in Github Pages, Click here
Navigate to the Neural_Network directory and run:: python train_network.py
To recognize digits from image use as:: python Recognize_Digit.py path/to/image path/to/outputfile
You can find the result in outputfile. In case the path to the output is not provided, the output is displayed on the console.
Recognising text has multiple applications including but not limited to Number Plate Recognition, Medical applications - scanning and documenting prescriptions, evaluating examinations etc.
This project has started as part of course on Software Development.
The pre-requisites for this software are
- Numpy
- Scipy
- PIL
- OpenCV
- Matplotlib
Install all the pre-requisites and run:: python setup.py install
Let people know how they can dive into the project, include important links to things like issue trackers, irc, twitter accounts if applicable.
Harish Murali, email-id: harish2111996@gmail.com
Surya Mohan, email-id: suryamohan1919@gmail.com, suryamohan@iitb.ac.in