Detecting objects, using region proposal-based CNN

In the previous section, we have learned about generating region proposals from an image. In this section, we will leverage the region proposals to come up with object detection and localization within an image.

Getting ready

The strategy we shall be adopting to perform region proposal-based object detection and localization is as follows:

1. For the current exercise, we'll build the model based on images that contain only one object
2. We'll extract the various region proposals (candidates) within the image
3. We will calculate how close the candidate is to the actual object location:
   • Essentially, we calculate the intersection over the union of the candidate with the actual...