Hands-On Java Deep Learning for Computer Vision

Resolving object localization

In this section, we'll look at an interesting solution to solving the object localization and detection problem. In addition, we'll learn how to label data and modify the prediction layer based on the need of the localization problem. We'll extend this to landmark detection, which should enable us to do fascinating things, such as detect a person smiling. 

We're familiar with the image classification problem; we need to determine whether the image contains any of the desired classes, such as whether the object is a car or a person. Apart from image recognition, we also want to find the position of the object or localize the object by marking it with a bounding box:

We'll also extend the idea of localization ...