/MNIST-Handwritten

Classification of MNIST handwritten digits & custom handwritten digits.

Primary LanguagePython

PROBLEM

Train a convolutional neural network to recognize the MNIST handwritten digits.

TODO

  1. Succesful training and testing of classifier indicating training and test accuracies.
  2. Live demo presenting capture and detection of handwritten digits.
  3. Extra Credit: Live Demo presenting localization and identification of a handwritten digit that has been embedded in another image.

Results

Training:

Loss: 0.0620, Acccuracy: 98%

Testing:

Loss: 0.03162, Accuracy: 98.91%