Image classification is a major application domain for artificial intelligence (AI) in the modern day. We can find instances of image classification in a large number of places all around us, such as face unlocking for mobile phones, object recognition, optical character recognition, tagging of people in photos, and several others. While these task seems pretty simple when you think of it from a human's perspective, it is not as simple when it comes to computers. Firstly, the system has to recognize objects or people from an image and draw a bounding box around it/them and proceed to classification. Both these steps are compute-intensive and hard to perform for machines.
There are several challenges in image processing that researchers are trying to overcome every day, such as face recognition for people with glasses on or a newly grown...