DigiTell is a python Program that is a basic replication of an Optical Character Recognition (OCR) which can be used to detect Hand Written digits live at camera or through a video.
Handwritten character recognition is one of the practically essential issues in pattern recognition applications. The applications of digit recognition include postal mail sorting, bank check processing, form data entry etc. The heart of the problem lies within the ability to develop an efficient algorithms that can recognize handwritten digits and which is submitted by users by way of a scanner, tablet, and other digital devices.
This Repository presents an approach to off-line handwritten digit recognition based on different machine learning techniques. The main objective of this repository is to ensure effective and reliable strategies for the Recognition of handwritten digits. This Repository Tries to replicate the Machine learning algorithm, namely Support Vector Machine (SVM) which has been used for the Recognition of the digits.
Run performRecognition live.py if you have a webcam and detect your images. Run performRecognition Scanned.py if you don't have a webcam and detect the scanned images. You can also upload your images of handwritten digits and specify them in the program.
We can Run perform Recognition scanned if you have scanned images with digits written in it or you can also use sample_1.jpg or sample_2.jpg for scanning.
This Repository uses the MNIST dataset of handwritten digits and then tries to train an optimized model that should be fast enough to work in realtime even on low computational devices with a high accuracy rate.