Detecting objects with YOLOv3
In the Creating an object detector with image pyramids and sliding windows recipe, we learned how to turn any image classifier into an object detector, by embedding it in a traditional framework that relies on image pyramids and sliding windows. However, we also learned that this approach isn't ideal because it doesn't allow the network to learn from its mistakes.
The reason why deep learning has conquered the field of object detection is due to its end-to-end approach. The network not only figures out how to classify an object, but also discovers how to produce the best bounding box possible to locate each element in the image.
On top of this, thanks to this end-to-end strategy, a network can detect a myriad objects in a single pass! Of course, this makes such object detectors incredibly efficient!
One of the seminal end-to-end object detectors is YOLO, and in this recipe, we'll learn how to detect objects with a pre-trained...
