Image recognition is a computer vision technique that searches the input image for a particular bitmap pattern. Our image recognition algorithm should be able to detect the pattern even if it is scaled, rotated, or has different brightness than of the original image.
How do we compare the pattern image against other images? As the pattern can be affected by perspective transformation, it's obvious that we can't directly compare pixels of the pattern and test image. The feature points and feature descriptors are helpful in this case. There is no universal or exact definition of what the feature is. The exact definition often depends on the problem or the type of application. Usually a feature is defined as an "interesting" part of an image, and features are used as a starting point for many computer vision algorithms. In this chapter we will use a feature point term, which is a part of the image defined by a center point, radius...
