This repository contains an example demonstration of how a convolutional neural network can be used to classify handwritten letters provided by the EMNIST dataset.
The EMNIST-letters dataset contains 145,600 handwriting samples balanced across 26 classes (one class per English letter). The purpose of this repository is to train a convolutional neural network to classify uppercase and lowercase handwritten letters from this dataset. Images of various clarity will be fed into the network to determine how image blurriness affects network learning.
This project uses tensorflow and extended keras datasets
- pip install tensorflow
- pip install extra-keras-datasets