https://github.com/burakahmet/handwriting-digit-recognition

Instant handwriting digit classification with mnist dataset and TensorFlow
https://github.com/burakahmet/handwriting-digit-recognition

artificial-intelligence artificial-neural-networks cnn colab-notebook convolutional-neural-networks data-augmentation handwritten-digit-recognition image-processing keras mnist model numpy pillow python tensorflow tensorflow-examples tensorflow-models tkinter tkinter-canvas tkinter-graphic-interface

Last synced: 1 day ago
JSON representation

Instant handwriting digit classification with mnist dataset and TensorFlow

• Host: GitHub
• URL: https://github.com/burakahmet/handwriting-digit-recognition
• Owner: BurakAhmet
• License: apache-2.0
• Created: 2024-02-08T19:40:58.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-02-12T10:20:14.000Z (10 months ago)
• Last Synced: 2024-11-03T06:42:25.801Z (about 2 months ago)
• Topics: artificial-intelligence, artificial-neural-networks, cnn, colab-notebook, convolutional-neural-networks, data-augmentation, handwritten-digit-recognition, image-processing, keras, mnist, model, numpy, pillow, python, tensorflow, tensorflow-examples, tensorflow-models, tkinter, tkinter-canvas, tkinter-graphic-interface
• Language: Jupyter Notebook
• Homepage:
• Size: 12.1 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Handwriting Digit Recognition

This project enables the recognition of handwritten digits using TensorFlow and Tkinter libraries with mnist dataset. After drawing a digit on the Tkinter interface, the TensorFlow model is used to predict the drawn digit.

I used **CNNs** (Convolutional Neural Networks) and **data augmentation** techniques to get high val-accuracy result.

## Preview

https://github.com/BurakAhmet/Handwriting-Digit-Recognition/assets/89780902/0e31bbf3-eae2-4f34-85c3-37c3b1cf0986

## Model

I used CNN (Convolutional Neural Networks) and data augmentation techniques in my model

### Model Accuracy

Final training loss: **0.0444**

Final training accuracy: **0.9858**

Final validation loss: **0.0182** 

Final validation accuracy: **0.9948**

![model accuracy](https://github.com/BurakAhmet/Hand-Writing-Digit-Recognition/assets/89780902/c2566e6c-ea26-4f98-b929-b43317bc8828)

## Technologies Used

* Python 3: The project is developed using Python programming language.

* Pillow (PIL): Utilized for capturing and processing images.

* TensorFlow: Used for training the data, loading pre-trained models and making predictions.

* NumPy: Employed for array manipulation and normalization of input data.

* Tkinter: Utilized for creating the user interface (canvas).

* Google Colab: Used for fast model training with GPUs.