SVM Classification with HOG Image Processing

Inspiration

I am a TA for a class on data analysis. I have been meaning to make a tutorial on data analysis for the students.

What it does

This code processes handwriting data (0's and 1's) from the MNIST dataset. It uses the Histogram Orient Gradient (HOG) method to process png images (after some pre-processing) and then performs SVM classification on them.

How I built it

I used handwriting images from the MNIST handwriting database. I put 0's into a folder called "Zero" and 1's into a folder called "One." I implemented the algorithm.

What I learned

How to write a decent tutorial. But, more importantly, that I need to manage my time better so that I am not doing academic work during hackathons.

What's next for SVM Classification with HOG Image Processing

I'm going to make it so that it classifies more than just 0's and 1's.

Built With

• r