Aims:
-
Use tensorflow to recognise handwritten digits
-
Use OpenCV to get a bounding box around handwritten digits in an image 2a) Use B-Human's algorithm to find the optimal threshold point, and binarise the image 2b) Dilate the black 2c) Find black regions, find the bounding box (watershed/floodfill) 2d) this gets a box around each digit
-
Classify these digits, and replace within the bounding box with a neatly typed digit, and also store and print the string. keeping formatting like new lines, etc
-
Long term goal - extend this to the alphabet, and be able to convert handwritten pages to text
Useful links: https://colah.github.io/posts/2014-10-Visualizing-MNIST/ http://docs.seldon.io/tensorflow-deep-mnist-example.html http://cv-tricks.com/tensorflow-tutorial/training-convolutional-neural-network-for-image-classification/ http://cv-tricks.com/tensorflow-tutorial/save-restore-tensorflow-models-quick-complete-tutorial/ https://www.oreilly.com/learning/not-another-mnist-tutorial-with-tensorflow
- tensorflow docs