Image features

In this section, we will understand what we mean by image features and why they are important.

Imagine a case where you see a person and you immediately detect the face of the person—his eyes, nose, and many other facial features. The question is how do we do this? What is the algorithm that we follow to detect these facial features? How do we describe them? Additionally, when we see another person with the same facial features, we can easily spot the matching features between the two persons. What is the metric that we use to measure this similarity?

We simply follow the process of detecting, describing, and matching the features. From the computer's perception, we want the process to be able to find features that can be repeatedly extracted, adequately represented, and accurately matched.

These features are considered good features, and to measure the goodness of a feature, we should consider its robustness and invariance (especially, invariance to scale and rotation...