/number-recognition-using-python

Can recognize the handwriting mostly(only) numbers with accuracy using ML & dataset with GUI to write numbers.

Primary LanguagePython

Number recognization using python

Recognize number with accuracy percentage in GUI

number-recognition-using-python

Features

  • Added dataset. No need to set any new training data.
  • Added dll file too. Not important as of now here.
  • Attached .mnist file for data retrieval.
  • Can recognize single integer numbers(0-9) only as of now. [single numbers]
  • Can display the accuracy of the result.
  • You can draw numbers with your mouse.

Modules used

Modules depend on the version of python you're using, as it's the env. Let me tell you my modules. Only versions will change.

  1. keras(most important for ML project)
  2. numpy (of course important)
  3. PIL (very interesting module)
  4. tkinter(as if you can make GUI without this)
  5. win32gui(I've used pywin32 btw)
  6. tensorflow(Also important)
  • NOTE:-Please check if I'm forgetting any module/library. I just forgot them all at once.

How to run

Steps to run this project

Step 1. Extract the repository
Step 2. Open this project with PyCharm Community(I use this)
Step 3. Install all above modules/libraries
Step 4. Run the file gui_digit_recognizer.py

Screenshots

  1. The error which I got while running in PyCharm after update (This is necessary for a programmer I guess) ERRORS are important

  2. This is the first GUI you'll see after running the file GUI made after ages

  3. This is the result you'll got(I've got, you'll also get I'm sure) recognize 2

recognize 4