Image classifier for vehicules found on EPFL campus - Swissroads

Objectives

Primary: Test different classifiers and techniques using high-level features extracted with a pretrained convolutional neural network from TensorFlow Hub and compare results with own ConvNet implementation trained fomr the raw image pixels
Secondary: Build an image classifier using Swissroads dataset which contains several hundreds images of vehicles found in the EPFL in Lausanne

Content

• Classifiers:
  • k-NN
  • Decisiont Tree
  • Logistic Regression
  • Random Forest
  • SVMs
  • Keras

For more details see GitHub ML projects - Swissroads - Gregoire LAURENT.