Once we have detected the interest points in an image, it would be good to know how to match the points across different images of the same object. For example, the following list shows the general approach to matching two such images:

• Compute the points of interest (for example, corner points with the Harris Corner Detector)
• Consider a region (window) around each of the key-points
• From the region, compute a local feature descriptor, for each key point for each image and normalize
• Match the local descriptors computed in two images (using Euclidean distance, for example)

Harris Corner points can be used to match two images; the next section gives an example.