This edited compendium of chapters represents the largest effort to date to bring together the breadth and depth of image processing research for document text extraction, segmentation of document image into picture and text zones, and general optical character recognition (OCR) of the international family of foreign languages. The result of this effort manifests itself in this two-tome handbook containing 30 chapters on a multitude of topics. The authors of these chapters wanted to convey their expertise on the corresponding topics in a comprehensive manner.
The editors are well established in the field of OCR and the related field of document image processing (DIP). Doermann recently joined DARPA after being elected as a Fellow of the IEEE for his contributed works on the automatic processing of documents. He authored more than 250 papers and presided over the Language and Media Processing Laboratory in College Park, Maryland. Tombre is currently a vice president of the University of Lorraine, France, and created the versatile graphics recognition library, QGAR.
A consistent format is used throughout the book to enhance the reading and learning experience. First, the chapters are divided into eight parts (fundamentals, page analysis, text extraction, general page processing, practical applications, online data processing, benchmarking, and evaluation) coalescing similar topics under one rubric. The largest part on text recognition is seven chapters long. The editors explain the cohesiveness in an introductory statement included at the beginning of each part.
Included in the prefaced remarks to each part by the editors is a brief statement about each chapter contained within written by respective experts in their fields. This provides a framework to understand the material that follows. The authors complied with the format chosen by the editors: the papers start with an introduction, followed by a clear presentation of the topic, and conclude with a summary of the significant points in the chapter with a meaningful list of references so that the reader can explore the topics further. The authors make an effort to provide the reader with information about existing software and corresponding datasets for testing purposes. A number of times, the chapters even make reference to each other in order to present a unified approach to the material.
Besides the more classic topics such as image acquisition (chapter 2), page segmentation (chapter 5), and logical layout (chapter 6), this handbook has a number of interesting and diverse topics. Font recognition is discussed in chapter 9 and handwriting analysis in chapters 11 (hand-printed characters), 12 (handwritten script), 26 (online handwriting), and 27 (signature verification). Specialized alphabets such as Middle Eastern and Asian are discussed in chapters 13 and 14, respectively; logo recognition is covered in chapter 18; mathematical notations are addressed in chapter 20; some advanced applications such as postal applications and check processing appear in chapter 21; and analyzing musical scores is discussed in chapter 22.
This handbook will appeal to the widest audience possible, including academicians, practitioners, library science and legal professionals, and all who are interested in the efficient storage and retrieval of vast numbers of documents.
