After learning about image classification in the last chapter, we now know that when we classify an image, we are just trying to output the class label of an object inside that image. Usually, to make the task easier, there will be only one object in the image.

Moving forward, in many cases, we are also interested in finding the location of an object in the image. The name given to this task of locating an object is called localization. In this case, the output we want to produce is the coordinates of a box that goes around the object. The name for this box is the bounding box or bounding rectangle. The important detail about localization is that we only localize one object per image.

When we build a model responsible for predicting a class label as well as the bounding box around the object of interest, it is called image classification with localization.