This is a simple real time object detection Android sample application, what uses TensorFlow Mobile to detect objects on the frames provided by the Camera2 API. In this example we are going to show you how it works with a tiny-yolo model. The demo application is able to detect twenty classes of the Pascal VOC dataset, for sure you can modify it by changing the ProtoBuff (*.pb) and label (*.txt) files in the /assets directory. Later we would like to port it to TensorFlow Lite to become a more optimal solution, also to add support for YOLOv3.
Bicycle detected with Android YOLO
Prerequirements:
Compile and run the project: