trEN

AI model

Plastic Waste Classification

An image classification model that tells bottles, bags and packaging apart. I collected the 4,500-image dataset myself; test accuracy is about 85–90%.

PythonResNet50CNN
  • done

What it does

A CNN-based image classification model that separates bottles, bags and packaging.

How

  • I collected the dataset myself: 1,500 images per class, 4,500 in total.
  • A ResNet50-based model, in Python.

Outcome

About 85–90% accuracy on the test set.