As optical character recognition (OCR) begins to find applications ranging from store checkout scanners to money-changing machines and postal system automation, it has become one of the most dynamic areas in information science today. Yet few volumes explore this data-oriented process without relying heavily on mathematical background reading. Now, Shunji Mori, Hirobumi Nishida, and Hiromitsu Yamada, among the field's most respected researchers since its inception, present this self-contained, clearly written guidebook to OCR--the first comprehensive treatment of the preprocessing, feature-extraction, and systematic description-matching stages of the OCR process. Including a wealth of original research material available here for the first time, this book is both an ideal professional reference source and an excellent entry point for course work in the subject. Key features of Optical Character Recognition: Theoretical framework based on functional analysis--not previously available in a detailed, English-language version Extensive explanation of preprocessing theory, including blurring and sampling, normalization, thinning, and binary and gray-scale morphology Intensive section on feature extraction, exploring linear methods, structure analysis, and algebraic description Original work on systematic shape description as a prerequisite to matching Original material on elastic matching, including image recognition of characters and objects Requires only the standard undergraduate requisites of algebra, linear algebra, and advanced calculus