This chapter is a brief introduction to the principles of automatic object recognition. It introduces the basic terms, concepts, and approaches to feature-based classification. It also introduces the term invariant and provides an overview of the invariants which have been proposed for visual object description and recognition. The chapter presents eight basic ones-simple shape features, complete visual features, transformation coefficient features, wavelet-based features, textural features, differential invariants, point set invariants, and moment invariants. It briefly reviews the basic classifier types and two popular techniques used for improving the classification-classifier fusion and dimensionality reduction. The nearest-neighbor (NN) classifier, sometimes also called the minimum distance classifier, is the most intuitive classifier. Classifiers called the Support vector machines (SVMs) are generalizations of a classical notion of linear classifiers. Artificial neural networks (ANNs) are ‘biologically inspired’ classifiers. The chapter explains how to increase the classifier performance, and shows how the classifier performance should be evaluated.