Skip to content

Classifying images from three types of lego sets with the help of a neural network built on the basis of ResNet50.

Notifications You must be signed in to change notification settings

Alessine/legoset_classification

Repository files navigation

Legoset classification

I built a convolutional neural network on the basis of ResNet50 with the goal of classifying images of used lego sets from three different product lines: Duplo, Star Wars and Technic. I am using the data generated with my Ricardo webscraping code (see my other repository).

The network currently achieves a classification accuracy of around 75%. It is work in progress. You can check out all the code in the Jupyter Notebook.

Note: This repository was inspired by an in-class exercise during my Data Science Bootcamp at Propulsion Academy.

alt text alt text alt text alt text alt text

About

Classifying images from three types of lego sets with the help of a neural network built on the basis of ResNet50.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published