Igor Susmelj • 2/10/2020

A random forest image classifier in a day

This article details a Hack Zurich 2016 project where the team created a sign digitizer. Faced with no existing dataset, they quickly drew and scanned signs, used OpenCV for cropping, and built a random forest classifier with scikit-learn on 16x16 pixel images, all within a single day. It covers data collection, preparation, model training, and integration into a Flask API.

0 comments

#Machine Learning #Opencv #Scikit Learn