Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.

Gradient-based learning applied to document recognition

Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

The primary goal of pattern recognition is supervised or unsupervised classification. Among the various frameworks in which pattern recognition has been traditionally formulated, the statistical approach has been most intensively studied and used in practice. More recently, neural network techniques and methods imported from statistical learning theory have been receiving increasing attention. The design of a recognition system requires careful attention to the following issues: definition of pattern classes, sensing environment, pattern representation, feature extraction and selection, cluster analysis, classifier design and learning, selection of training and test samples, and performance evaluation. In spite of almost 50 years of research and development in this field, the general problem of recognizing complex patterns with arbitrary orientation, location, and scale remains unsolved. New and emerging applications, such as data mining, web searching, retrieval of multimedia data, face recognition, and cursive handwriting recognition, require robust and efficient pattern recognition techniques. The objective of this review paper is to summarize and compare some of the well-known methods used in various stages of a pattern recognition system and identify research topics and applications which are at the forefront of this exciting and challenging field.

/pdf/statistical-pattern-recognition-a-review-1kz1b0tte4.pdf

Statistical pattern recognition: a review

We develop a common theoretical framework for combining classifiers which use distinct pattern representations and show that many existing schemes can be considered as special cases of compound classification where all the pattern representations are used jointly to make a decision. An experimental comparison of various classifier combination schemes demonstrates that the combination rule developed under the most restrictive assumptions-the sum rule-outperforms other classifier combinations schemes. A sensitivity analysis of the various schemes to estimation errors is carried out to show that this finding can be justified theoretically.

On combining classifiers

From the Publisher:
Pattern recognition has long been studied in relation to many different (and mainly unrelated) applications, such as remote sensing, computer vision, space research, and medical imaging. In this book Professor Ripley brings together two crucial ideas in pattern recognition; statistical methods and machine learning via neural networks. Unifying principles are brought to the fore, and the author gives an overview of the state of the subject. Many examples are included to illustrate real problems in pattern recognition and how to overcome them.This is a self-contained account, ideal both as an introduction for non-specialists readers, and also as a handbook for the more expert reader.

Pattern recognition and neural networks

Whilst the design of skeletonization algorithms has been a very active research area, methodologies for an automatic evaluation of the quality of the results remain to be developed. The difficulty rests on the fact that certain geometric properties considered desirable in skeletons (especially for pattern recognition applications) are not easily quantifiable by means other than human vision. The attempt here is to develop distance measures based on shape matching methods, and to compare skeletons to references in terms of these distances. Three such methods have been adapted for this purpose, and the results are found to be highly correlated for the samples tested.

Automatic comparison of skeletons by shape matching methods

Previous research efforts on optical font recognition have mostly limited applications since they deal with only a few types of font attributes and estimate them from a line or block of text This paper proposes a word-level optical font recognition system for printed Korean and English documents At the word-level, it has the advantages of obtaining more detailed font attributes including the following: script (Korean and English), font style (regular, bold, italic, and underlined), typeface (Myung-jo and Gothic), point size (10, 12, 14 pts), and word length (2, 3, 4, 5 for Korean, and 4 to 10 for English) A hierarchical classifier and several typographical features have been devised for the system, and their effectiveness are proven by an experiment with a database of 100 sets of 264 font categories

Word-level optical font recognition using typographical features

Presents an approach to differentiating handwritten alphabetic words from numeric strings, which is motivated by an application in a real-life bank cheque processing system. Neural networks are first considered with the focus on effective feature extraction and architecture selection. The combination of networks is then implemented to achieve highly reliable and accurate results.

Differentiation between alphabetic and numeric data using NN ensembles

The Histogram of Oriented Gradients (HOG) is one of the most widely used methods to extract the gradient features for object recognition and consistently shows high accuracy rates when compared to other descriptors. The major drawbacks of using the HOG method are the necessity of finding an optimal window size to fit the whole object; and the exhaustive search mechanism represented by a fixed window size sliding through the whole image to locate and recognize objects. This research proposes an efficient and robust Dynamic-HOG as an improvement to the traditional HOG method to locate and recognize structured objects in images. The proposed method works by locating and analyzing the structured objects in images in order to define a dynamic window size w.r.t. each object size. Moreover, the Dynamic-HOG method requires much less processing time by eliminating the exhaustive search mechanism. The method defines the height and width thresholds of objects and bounds each object with a window w.r.t. its size while ignoring non–object edges. It fits structured objects of a close range of heights and widths. This paper considers the characters that are minted on coins of different languages and sizes as the objects to recognize. There are several papers in the literature discussing coin recognition problem and proposing solutions based on various sets of features extracted from the entire coin image. This research also proposes a new method for coin recognition by focusing on recognize coins based on smaller part of the coin image which are the characters. Our method is evaluated on coins from diverse countries with different background complexity. The proposed method achieved precision and recall rates as high as 98.08 and 98.23%, respectively; which demonstrate the effectiveness and robustness of the proposed method.

Ching Y. Suen

Papers

Automatic comparison of skeletons by shape matching methods

Word-level optical font recognition using typographical features

Differentiation between alphabetic and numeric data using NN ensembles

An Enhanced Histogram of Oriented Gradient Descriptor for Numismatic Applications

Word spotting techniques in document analysis and retrieval — a comprehensive survey