Provides a foundational review and perspective of grammatical approaches to computer vision. In its quest for a stochastic and context sensitive grammar of images, this book is intended to serve as a unified frame-work of representation, learning, and recognition for a large number of object categories.