Feature Extraction and Image Processing for Computer Vision, 4th Edition

5.4. Feature extraction by low-level features

There have been many approaches to feature extraction which combine a variety of features. It is possible to characterise objects by measures that we have already developed, by low-level features local features (such edges and corners), and by global features (such as colour). Later we shall find these can be grouped to give structure or shape (in this chapter and the next), and appearance (called texture, Chapter 8). The drivers for the earlier approaches which combine low-level features are the need to be able to search databases for particular images. This is known as image retrieval (Section 4.4.2), and in content-based retrieval that uses techniques from image processing and computer vision, ...