Object Detection via pre-trained YOLOv3 is used to detect vehicles from an image. Transfer learning on ResNet-50 outer layers with a two-class dataset trains for 50 epochs. 81% accuracy.