DarkAdin / DigitRecognizer

A computer vision algorithm implemented to return a very acceptable submission for the challenge 'Digit Recognizer' from Kaggle: https://www.kaggle.com/competitions/digit-recognizer

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Digit Recognizer

A Python Jupyter notebook with a simple computer vision implementation for creating a very acceptable submission for the Digit Recognizer challenge in Kaggle: https://www.kaggle.com/competitions/digit-recognizer

The image above shows:

  • The digit number 4777
  • A slightly rotated version of itself
  • A slightly zoomed version of itself

The algorithm performs these augmentations for every digit in the dataset, effectively increasing its size to three times the original. The labels are triplicated to compensate for this.

The implemented algorithm returned a testing accuracy of $0.986380934715271$ as is, but the performance can be easily improved.

In the image above, both the training/test accuracy and error can be seen for three epochs, generated by the notebook as is.

About

A computer vision algorithm implemented to return a very acceptable submission for the challenge 'Digit Recognizer' from Kaggle: https://www.kaggle.com/competitions/digit-recognizer


Languages

Language:Jupyter Notebook 100.0%