Young-Bin Kwon

Line removal and restoration of handwritten characters on the form documents

The off-line handwritten characters recorded on prescribed form documents may be overwritten by the lines of the form documents. Overwritten characters should be isolated in order to be recognized more effectively. However, removal of the lines causes breaks in the overwritten characters. Consequently, a character restoration process is necessary. In this paper, the shape types of overwritten characters are analyzed and a method of restoring characters that have been broken by line removal is proposed. A 97% correct restoration ratio was obtained through this method.

Graphics Recognition Algorithms and Applications

Engineering solutions are generally documented in assembly and part drawings and bill of materials. A great benefit, qualitatively and commercially, can be achieved if these paper based storages can be transformed into digital information archives. The process of this transformation is called reconstruction. The reconstruction process of paper based assembly drawings consists of four steps: digitization; vectorization/interpretation; 3D reconstruction of the parts and the 3D reconstruction of the assembly. This paper evaluates existing commercial systems worldwide for interpretation of paper based mechanical engineering drawings. For a complete reconstruction process a 3D reconstruction is needed. This functionality is already supported by some CAD systems to a certain extent, but it still remains a major topic of research work. One CAD system which converts 2D CAD models into 3D CAD models is presented. Finally, after the reconstruction of the parts the whole assembly can be reconstructed. Until now, no system for the automatic reconstruction of assemblies is available. In our paper we present a general approach for automatic reconstruction of 3D assembly model data by interpretation of mechanical engineering 2D assembly drawings, their part drawings, and the bill of materials.

Graphics Recognition. Recent Advances and Perspectives

In this paper, we present different strategies for localization and recognition of graphical entities in line drawings. Most systems include first a segmentation step of the document followed by a sequential extraction of the graphical entities. Some other systems try to recognize symbols directly on the bitmap image using more or less sophisticated techniques. In our system, an intermediate representation of the document provides a precise description of all the shapes present in the initial image. Thereafter, this representation constitutes the main part of a shared resource that will be used by different processes achieving the interpretation of the drawings. The actions (recognition) done by these different specialists are scheduled in order to read and understand the content of the document. The knowledge that is provided by the shared representation is used instead of the bitmap image material to drive the interpretation process. In the current system, the specialists are trying, during several cycles to interpret the drawings in an intelligent way by interpreting the simplest parts of a drawing first and making the shared representation evolve until the total understanding of the document.

Main Wall Recognition of Architectural Drawings Using Dimension Extension Line

This paper deals with plain figures on the architectural drawings of apartment. This kind of architectural drawings consist of main walls represented by two parallel bold lines, symbols(door, window, tile...), dimension line, extension line, and dimensions represent various numerical values and characters. This paper suggests a method for recognizing main wall which is a back-bone of apartment in an architectural drawing. In this thesis, the following modules are realized: an efficient image binarization, a removal of thin lines, a vectorization of detected lines, a region bounding for main walls, a calculation of extension lines, a finding main walls based on extension line, and a field expansion by searching other main walls which are linked with the detected main walls. Although the windows between main walls are not represented as main walls, a detection module for the windows is considered during the recognition period. So the windows are found as a part of main wall. An experimental result on 9 different architectural drawings shows 96.5% recognition of main walls and windows, which is about 5.8% higher than that of Karl Tombre.

Changes to vascular biometric system security & performance

When considering biometrics for a personal identification system, different modalities can be considered. The final selection will depend on specific application requirements. From the several modalities existing nowadays, vascular systems have appeared on the scene lately. Vein-based identification is claimed to be as reliable as fingerprint or iris identification, but with the usability of hand geometry or even the face. As with any other biometric modalities, vascular solutions have to be deeply analyzed for all relevant factors that could affect their performance or the security level achieved. This will show the analysis of the performance and security achieved by a commercial vascular biometric system, when being applied in several real-world scenarios. These scenarios will cover typical office environments to the more extreme environments, such as extreme light and temperatures, as those suffered in a banking ATM, or with high humidity such as in a gym. The results will show the strengths of this modality, as well as those points where further improvements are needed.

Information Extraction from a Skewed Form Document in the Presence of Crossing Characters

In this paper, we propose an information extraction method which can restore the handwritten character information from prescribed and skewed form documents. The proposed method include the following aspects: boundary and successive internal line dectection, accurate skew angle measurement, line removal and broken character restoration using morphological analysis model of crossing shape. Using the proposed method, more than 95% of the horizontal and vertical crossing lines are correctly restored.

Automatic Region Labeling of the Layered Map

In this paper, we describe an automatic region labeling method, which identifies each region and recognizes region names. Before tracing the region boundaries, it extracts the region names which consist of characters, dots, dashes, and indication lines. It uses two recognition methods to recognize characters in the region name. In the case of recognizing the isolated characters, it uses the open and close features. The characters touching boundaries are recognized by template matching. After removing the components of region names from a map image, the boundaries of each region are extracted. After which it then vectorizes the region boundaries. From these recognition results, the original map can be constructed. It reduces the storage to one fifth of the original data. The proposed method shows 95% accuracy of region labeling.

Standardised system for automatic remote evaluation of biometric algorithms

It is difficult to generate significant results when evaluating algorithms developed in biometrics. Additionally, when third parties, such as system integrators, want to compare results among different algorithm providers, they cannot do so easily for several reasons: difficulty accessing large databases, inability to exchange biometric data among researchers due to data protection laws in some countries, or the lack comprehensive and standardised evaluation reports. This paper presents a new performance evaluation system for biometric systems that is secure, automatic and remote. This system has been developed using current standards developed within ISO/IEC JTC1/SC37 for data Formats, Application Program Interfaces (APIs) and evaluation methodology. Standardised technology is able to provide developers in biometrics and third parties with a way to perform comprehensive evaluations remotely and with 24/7 availability without compromising the privacy of the individuals included in the test crew. The solution described here offers the developers the ability to evaluate large databases that are stored in a secured centralised server. As this system is modality-independent, researchers can use the same protocol to perform different evaluations, and therefore lower the overhead costs for testing purposes. Additionally, such protocols can be plugged directly into end-user applications, minimising technology transfer costs. The system is described by block diagrams as well as flowcharts.

Implementation of Content Analysis System for Recognition of Journals_ Table of Contents

In this paper, we design the primary component, an automated segmentation method that extracts article titles, author names and page numbers from scanned image of a journals table of contents. We expanded the method to five types of journal layout. An interactive tool was also developed with Visual C++ to help the user manage the segmentation results, and manually control the table analysis. Its performance was tested over a panel of seventy images, which gather the most common types of layout encountered.

Multi-font and multi-size character recognition based on the sampling and quantization of an unwrapped contour

We propose a character recognition method based on the sampling and quantization of unwrapped contour information. The character contour includes many different features. Numerous approaches are explored to find essential features which are extracted directly from the contour. Our idea is to decompose the multi-dimensional features of a contour into one dimensional features. Following the contour, we extract horizontal, vertical, and angular variations and make unwrapped contours from these features. Unwrapped contour information is sensitive to the starting point of contour tracing, and it varies on the font type and size, so we propose a novel algorithm which finds the invariant starting point of contour tracing. By sampling and quantization, we can make discrete contours which are invariant to the font and size.