George Nagy

Twenty years of document image analysis in PAMI

The contributions to document image analysis of 99 papers published in the IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI) are clustered, summarized, interpolated, interpreted, and evaluated.

A prototype document image analysis system for technical journals

Gobbledoc, a system providing remote access to stored documents, which is based on syntactic document analysis and optical character recognition (OCR), is discussed. In Gobbledoc, image processing, document analysis, and OCR operations take place in batch mode when the documents are acquired. The document image acquisition process and the knowledge base that must be entered into the system to process a family of page images are described. The process by which the X-Y tree data structure converts a 2-D page-segmentation problem into a series of 1-D string-parsing problems that can be tackled using conventional compiler tools is also described. Syntactic analysis is used in Gobbledoc to divide each page into labeled rectangular blocks. Blocks labeled text are converted by OCR to obtain a secondary (ASCII) document representation. Since such symbolic files are better suited for computerized search than for human access to the document content and because too many visual layout clues are lost in the OCR process (including some special characters), Gobbledoc preserves the original block images for human browsing. Storage, networking, and display issues specific to document images are also discussed.<<ETX>>

A Comparative Study of Local Matching Approach for Face Recognition

In contrast to holistic methods, local matching methods extract facial features from different levels of locality and quantify them precisely. To determine how they can be best used for face recognition, we conducted a comprehensive comparative study at each step of the local matching process. The conclusions from our experiments include: (1) additional evidence that Gabor features are effective local feature representations and are robust to illumination changes; (2) discrimination based only on a small portion of the face area is surprisingly good; (3) the configuration of facial components does contain rich discriminating information and comparing corresponding local regions utilizes shape features more effectively than comparing corresponding facial components; (4) spatial multiresolution analysis leads to better classification performance; (5) combining local regions with Borda count classifier combination method alleviates the curse of dimensionality. We implemented a complete face recognition system by integrating the best option of each step. Without training, illumination compensation and without any parameter tuning, it achieves superior performance on every category of the FERET test: near perfect classification accuracy (99.5%) on pictures taken on the same day regardless of indoor illumination variations, and significantly better than any other reported performance on pictures taken several days to more than a year apart. The most significant experiments were repeated on the AR database, with similar results.

State of the art in pattern recognition

This paper reviews statistical, adaptive, and heuristic techniques used in laboratory investigations of pattern recognition problems. The discussion includes correlation methods, discriminant analysis, maximum likelihood decisions minimax techniques, perceptron-like algorithms, feature extraction, preprocessing, clustering and nonsupervised learning. Two-dimensional distributions are used to illustrate the properties of the various procedures. Several experimental projects, representative of prospective applications, are also described.

Rapid automated three-dimensional tracing of neurons from confocal image stacks

Algorithms are presented for fully automatic three-dimensional (3D) tracing of neurons that are imaged by fluorescence confocal microscopy. Unlike previous voxel-based skeletonization methods, the present approach works by recursively following the neuronal topology, using a set of 4 /spl times/ N/sup 2/ directional kernels (e.g., N = 32), guided by a generalized 3D cylinder model. This method extends our prior work on exploratory tracing of retinal vasculature to 3D space. Since the centerlines are of primary interest, the 3D extension can be accomplished by four rather than six sets of kernels. Additional modifications, such as dynamic adaptation of the correlation kernels, and adaptive step size estimation, were introduced for achieving robustness to photon noise, varying contrast, and apparent discontinuity and/or hollowness of structures. The end product is a labeling of all somas present, graph-theoretic representations of all dendritic/axonal structures, and image statistics such as soma volume and centroid, soma interconnectivity, the longest branch, and lengths of all graph branches originating from a soma. This method is able to work directly with unprocessed confocal images, without expensive deconvolution or other preprocessing. It is much faster that skeletonization, typically consuming less than a minute to trace a 70 MB image on a 500 MHz computer. These properties make it attractive for large-scale automated tissue studies that require rapid on-line image analysis, such as high-throughput neurobiology/angiogenesis assays, and initiatives such as the Human Brain Project.

Syntactic segmentation and labeling of digitized pages from technical journals

A method for extracting alternating horizontal and vertical projection profiles are from nested sub-blocks of scanned page images of technical documents is discussed. The thresholded profile strings are parsed using the compiler utilities Lex and Yacc. The significant document components are demarcated and identified by the recursive application of block grammars. Backtracking for error recovery and branch and bound for maximum-area labeling are implemented with Unix Shell programs. Results of the segmentation and labeling process are stored in a labeled x-y tree. It is shown that families of technical documents that share the same layout conventions can be readily analyzed. Results from experiments in which more than 20 types of document entities were identified in sample pages from two journals are presented. >

Geographic Data Processing

This survey attempts to provide a umfied framework for the constituent elements-originating In numerous and diverse disciplines--of geographical data processing systems. External aspects of such systems, as perceived by potential users, are discussed with regard to extent, coordinate system and base map, range of applications, input /output mechanisms, computer configuration, command and interaction, documentation, and administration.The internal aspects, which would concern the system designer, are analyzed in terms of the type of spatial variables involved and of their interrelationship with respect to common operations This point of view is shown to lead to a workable classificatmn of two-dimensional geometric algorithms and data structures. To provide concrete examples, ten representative geographic data processing systems, ranging from automated cartography to interactive decision support, are described. In conclusmn, some comparisons are drawn between geographmal data processing systems and their conventional business-oriented counterparts.

Towards Ontology Generation from Tables

At the heart of todays information-explosion problems are issues involving semantics, mutual understanding, concept matching, and interoperability. Ontologies and the Semantic Web are offered as a potential solution, but creating ontologies for real-world knowledge is nontrivial. If we could automate the process, we could significantly improve our chances of making the Semantic Web a reality. While understanding natural language is difficult, tables and other structured information make it easier to interpret new items and relations. In this paper we introduce an approach to generating ontologies based on table analysis. We thus call our approach TANGO (Table ANalysis for Generating Ontologies). Based on conceptual modeling extraction techniques, TANGO attempts to (i) understand a tables structure and conceptual content; (ii) discover the constraints that hold between concepts extracted from the table; (iii) match the recognized concepts with ones from a more general specification of related concepts; and (iv) merge the resulting structure with other similar knowledge representations. TANGO is thus a formalized method of processing the format and content of tables that can serve to incrementally build a relevant reusable conceptual ontology.

A Means for Achieving a High Degree of Compaction on Scan-Digitized Printed Text

A method of video compaction based on transmitting only the first instance of each class of digitized patterns is shown to yield a compaction ratio of 16: 1 on a short passage of text from the IEEE SPECTRUM. Refinements to extend the bandwidth reduction to 40: 1 by relatively simple means are proposed but not demonstrated.

Validation of image defect models for optical character recognition

Considers the problem of evaluating character image generators that model distortions encountered in optical character recognition (OCR). While a number of such defect models have been proposed, the contention that they produce the desired result is typically argued in an ad hoc and informal way. The authors introduce a rigorous and more pragmatic definition of when a model is accurate: they say a defect model is validated if the OCR errors induced by the model are indistinguishable from the errors encountered when using real scanned documents. The authors describe four measures to quantify this similarity, and compare and contrast them using over ten million scanned and synthesized characters in three fonts. The measures differentiate effectively between different fonts and different scans of the same font regardless of the underlying text.