Object detection and localization

Andrew Ng’s class on deep learning with convolutional neural nets covers object detection and localization with the YOLO algorithm, non-max suppression and anchor boxes.

Using the pretrained network from the YOLOv2 model, I picked a random street scene image and it did this:

References:

• You Only Look Once: Unified, Real-Time Object Detection (2015)

• YOLO9000: Better, Faster, Stronger (2016)

• YAD2K: Yet Another Darknet 2 Keras

• The official YOLO website