ColemanKobe/tensorflow-car-labeler

Using a Tensorflow Lite model, an Android application uses the camera and labels cars that it sees by model.

Table of Contents

• Summary
• Data Collection
• Testing
• Contribution

Summary

This Android application, based on Google's ML Kit, uses a TensorFlow model built on the basis of a collection of images. It detects and tries to label cars, and these labels comprise car models from different manufacturers.

Data Collection

All of these images are Creative Commons images, and were manually downloaded and collected in the cars directory.

Testing

This application is best tested using Android Studio, and this link provides information on how to run the application itself, and this link provides sufficient information to test the application on an Android mobile device.

Contribution

If you want to contribute to this application, it can be in two ways: adding data or improving the way the model is made in modelcreate.py. This should be done using the two branches that pertain to each way of contribution.