Krishna S. Nathan

On-line handwriting recognition using continuous parameter hidden Markov models

The problem of the automatic recognition of handwritten text is addressed. The text to be recognized is captured online and the temporal sequence of the data is presented. The approach is based on a left-to-right hidden markov model (HMM) for each character that models the dynamics of the written script. A mixture of Gaussian distributions is used to represent the output probabilities at each arc of the HMM. Several strategies for reestimating the model parameters are discussed. Experiments show that this approach results in significant decreases in error rate for the recognition of discretely written characters compared with elastic matching techniques. The HMM outperforms the elastic matching technique for both writer-dependent and writer-independent recognition tasks.<<ETX>>

A fast statistical mixture algorithm for on-line handwriting recognition

The automatic recognition of online handwriting is considered from an information theoretic viewpoint. Emphasis is placed on the recognition of unconstrained handwriting, a general combination of cursively written word fragments and discretely written characters. Existing recognition algorithms, such as elastic matching, are severely challenged by the variability inherent in unconstrained handwriting. This motivates the development of a probabilistic framework suitable for the derivation of a fast statistical mixture algorithm. This algorithm exhibits about the same degree of complexity as elastic matching, while being more flexible and potentially more robust. The approach relies on a novel front-end processor that, unlike conventional character or stroke-based processing, articulates around a small elementary unit of handwriting called a frame. The algorithm is based on (1) producing feature vectors representing each frame in one (or several) feature spaces, (2) Gaussian K-means clustering in these spaces, and (3) mixture modeling, taking into account the contributions of all relevant clusters in each space. The approach is illustrated by a simple task involving an 81-character alphabet. Both writer-dependent and writer-independent recognition results are found to be competitive with their elastic matching counterparts. >

Writer dependent recognition of on-line unconstrained handwriting

In this paper, we present a framework for adapting a writer independent system to a user from samples of the users writing. The writer independent system is modeled using hidden Markov models. Training for a writer involves recomputing the topology and parameters of the hidden Markov models using the writers data. The framework uses the writer independent system to get an initial alignment of the writers data. The system described reduces the error rate by an average of 65%. For the results presented, no language model was used.

Initialization of hidden Markov models for unconstrained on-line handwriting recognition

In a hidden Markov model system, the initialization of the model parameters is critical to the performance of the model after retraining. This paper proposes a number of new approaches to the problem of initialization, and demonstrates that a method of smooth alignment results in the best performance.

Size normalization in on-line unconstrained handwriting recognition

In an on-line handwriting recognition system, the motion of the tip of the stylus (pen) is sampled at equal time intervals using a digitizer tablet and the sampled points are passed to a computer which performs the handwriting recognition. In most cases, the basic recognition algorithm performs best for a nominal size of writing as well as a standard orientation (normally horizontal) and a nominal slant (normally fully upright). We discuss and provide solutions to these normalization problems in the context of on-line handwriting recognition. Most of the results presented are also valid for optical character recognition (OCR). Error rate reductions of 54.3% and 35.8% were obtained for the writer-dependent and writer-independent samples through using the proposed normalization scheme.<<ETX>>

Supervised hidden Markov modeling for on-line handwriting recognition

The performance of a large alphabet handwriting recognition system based on a probabilistic framework is critically tied to the quality of the prototype distributions that are established in the relevant feature space. To better account for handwriting variability, we describe a supervised strategy for the construction of prototype distributions which are more robust to allograph deformations. The idea is to incorporate supervision to relate the allographic models to their manifestations in the feature space. This makes for a better utilization of the available training data, while at the same time allowing for a short design time turn around. The performance of this method is illustrated on a discrete handwriting recognition task with an alphabet of 81 characters.<<ETX>>

The Clarion Call for modern services: China, Japan, Europe, and the U.S.

hat will modern services be like? Today many services are viewed as a craft activity—individual doctors, retail sellers, programmers all doing useful things their own way. There is, however, an increasing role for an organized, analytic, and engineering approach to all these activities. Evidence-based medicine, marketing sciencedriven retailers, and software engineering are examples of these trends. Automated services are a natural object of attention, since they can be observed in great detail, can be reconstructed and improved, and can be combined in new ways quickly and relatively easily. We are therefore seeing a rapid evolution toward an engineering approach to the life cycle of such services, and the application of mathematic and scientific approaches to the problems and opportunities they present. Complex service systems must be viewed at three levels: the functional attributes (what does it do and how does it do it?), nonfunctional attributes (management and control properties such as performance and security), and intentional attributes (what is the goal or purpose of the activity, such as societal benefit, private profit, or personal esteem?). Each level is susceptible to analysis, but different disciplines dominate. As computational services proliferate, new fields of study will open up, combining the computing, engineering, mathematical, management, and social sciences in creative ways. When we look at complex B2B projects, there is a growing application of solution engineering—using the best available techniques to the multiple phases of the activity, managing the risks, increasing predictability of quality and schedule, learning from experience in a project to improve not only the results of that effort but of succeeding solutions. As we examine the stages of a single large business service project (including requirements, design, implementation, deployment, and ongoing operation), and build up portfolios and service lines, much of the work can be formalized and subjected to analysis and radical improvement through optimization, evolutionary learning, and organization improvement. Of course, applying engineering thinking to such projects is not new—without such we would not have fields with names like “civil engineering” or large facilities like airports and suspension bridges. But the confluence of information-dominated services, techniques of computer science, and increasing experience is rapidly opening up new possibilities for modern service. A GLOBAL APPROACH Innovation is imperative to continued growth, increased productivity, and the general health of all economies. After a long history of contributions and breakthroughs to IT innovation, IBM Research is now directing resources toward innovation for the services industry: Business Design and Implementation. How does one model, design, and instantiate optimal business functions? What tools and techniques are needed to create abstractions of an enterprise, to effect the transformation from strategy down to underlying IT systems, and to monitor the end-to-end process? The Component Business Model (CBM) is one approach under development. What is needed to build and deliver industry solutions in an efficient, reusable fashion? Business Optimization and Management. How does one improve decision making and operations of ongoing business functions? How can business data be collected, analyzed, and exploited more optimally? How can business performance be enhanced as a result of the optimization of the underlying function, for example, supply chain? Given the importance of the work force in labor intensive services, how does one forecast, hire, allocate or shift resources to meet changing demand patterns? Services Delivery. In IBM’s internal service delivery centers, what are the best ways to maintain desired service levels while increasing efficiency and productivity? What tools and techniques can guarantee end-to-end manageability and visibility throughout the entire services life cycle from request to delivery? Challenges include the globalization of service delivery, adoption of standardized best practices, automation, virtualization of resources including labor, and the appropriate integration of the human element. Services Sciences Management and Engineering (SSME) applies to each of these areas [94]. IBM Research laboratories worldwide are engaging with local universities and governments on the topic of innovation in services, as described here.

Duration modeling results for an on-line handwriting recognizer

This paper describes a series of experiments that have been conducted to investigate the effect of duration modeling in a hidden Markov model (HMM) based online handwriting recognition system. The issues discussed include parametric vs. non-parametric distributions to model duration, different methods for training transition probabilities and the effect of weighting on duration terms.

A discrete parameter HMM approach to on-line handwriting recognition

One area where on-line handwriting recognition technology is most critical is the domain of small portable platforms. Because such platforms have limited resources, it is not presently practical to consider a continuous parameterization for the hidden Markov models used in the recognition. On the other hand, discrete parameter techniques such as used in speech recognition are difficult to apply, because there is no well-understood handwriting equivalent to phonological rules. A possible solution is to extract this information directly from the data, by constructing an alphabet of sub-character, elementary handwriting units. The performance of this method is illustrated on a discrete handwriting recognition task with an alphabet of 81 characters.

Parameter tying in writer-dependent recognition of on-line handwriting

We describe experiments on a writer dependent large vocabulary (20000) handwriting recognition system. The goal was to investigate the effects of different degrees of tying (and hence, of the number) of parameters on the error rate. The system recognizes cursive, printed or any combination thereof of script in real time on small PC platforms. Baseline results for a writer independent system are also included. Since we are only interested in the shape models no language models were used.