Organizing data involves image classification and feature extraction. In an image, there are some extra sets of features, which might not be of any significance to us. These extra features may be part of the background in an image or may be noise generated while converting the image into digital format. So, an image is made up of informative and important patterns combined with insignificant patterns.
In the image classification process, we extract the important and required features and leave out the rest. important features, we suppress and normalize the insignificant features and highlight the important features in an image. To extract these features, we use edge detection. Edge detection is a technique to find out the boundaries of an object in an image by detecting discontinuities in brightness. The image pixels of the object and the rest of the image background differ greatly in intensity and color, which forms the edge.

There are some well-known feature extraction techniques, namely HAAR-like features, Histogram of Oriented Gradients (HOG), Scale Invariant Feature Transform (SIFT), and Speed Up Robust Feature (SURF).