Tsuyoshi Shinogi

A segmentation method for touching Japanese handwritten characters based on connecting condition of lines

In unconstrained Japanese handwritten character strings, there are many touching characters. Segmentation of touching characters is required as preprocessing of isolated character recognition. But conventional segmentation methods cannot segment complicated touching characters. In this paper we propose a new segmentation method based on connecting condition of lines at a touching point, and evaluate the efficiency of this method for touching Japanese handwritten characters. This method could segment complicated touching characters with less unnecessary segmentations.

A simple and efficient method for generating compact IDDQ test set for bridging faults

This paper presents an efficient method for generating a compact test set for IDDQ testing of bridging faults in combinational CMOS circuits. This method is based on the iterative improvement method. Though our method is simple and easy to implement, it is efficient. Experimental results for large ISCAS benchmark circuits demonstrate its efficiency in comparison with results of previous methods.

Video Scene Segmentation Using the State Recognition of Blackboard for Blended Learning

We are developing the automatic generating system for video contents with keyword tag in blended learning. The generated lecture movie is too long for students to watch the lecture video on demand. We are considering the segmentation method of the lecture movie by the behavior of a lecturer. In this paper, we present a new segmentation method for one lecture movie to some shots using the behavior of a lecturer such as up-down movement of a blackboard and the erasure of characters on the blackboard. These behaviors are the sign of contents segmentation from the lecturer to students. The system detects the event time of these behaviors, and segments the lecture movie using the event time. We implemented our procedure and evaluated the validity using the lecture movie of some real lectures, and confirm that an accurate segmentation rate is 97%.

Region segmentation for table image with unknown complex structure

In this paper, we describe a system of region segmentation and conversion into an HTML file for an unknown machine-printed table image. Ruled lines delimit some cells of the table, and omitted ruled lines also delimit other cells. We consider a table analysis system for both types of table cell. First, our system segments a table by means of the ruled lines into some regions. Secondly, these segmented regions are further segmented into cells by the omitted ruled lines that are indicators (such as numerals and characters). The cells include several character lines, and our system can convert a table of unknown complex structure into an HTML file. Also, we confirm the effectiveness of our region segmentation method for various kinds of tables with omitted ruled lines by computer experiments.

Tendency discovery from incident report map generated by self organizing map and its development

This study discusses a tendency discovery method from medical text data with free format. In this study, we focuses incident reports and proposes a new tendency discovery method for them using self-organizing map. In the first of this paper, we describes the outline of the keyword extraction method from incident reports and coding method to make a map. It is expected that the generated map by the proposed method shows the relation among the incident reports visually and it gives us new knowledge, i.e. trends or features underlaid in reports and difficult to find from each report. The knowledge will be able to help the reduction of incidents in the hospital. Moreover, authors developed a incident report analysis system using the proposed method and discuss the effectiveness of the proposed system. Finally, the paper describes a possibility for the application of SOM to incident reports in the end of this paper.

A camera control based on fuzzy behavior recognition of lecturer for distance lecture

We describe a distance lecture support system using two cameras, which are a fixed view camera for image processing, and an active view camera for remote students, and propose a fuzzy camera control based on the behavior recognition of a lecturer. First, the silhouette of the lecturer is extracted from the input image of a fixed view camera. The parameters for behavior recognition of the lecturer such as face, body and hand are determined from the lecturers silhouette. A fuzzy reasoning system recognizes the lecturers behavior using the parameters. The active view area such as the portrait of the lecturer and the character strings on the blackboard is controlled by the fuzzy system automatically. The accuracy of the behavior recognition (70.4% for fuzzy reasoning and 55.2% for crisp reasoning) was obtained for the scenes of five lecturers.

Keywords Recognition of Handwritten Character String on Whiteboard Using Word Dictionary for e-Learning

We are developing an individual e-learning system using two communication cameras and a pen capture tool on whiteboard for university students. In this research, keywords recognition for the written characters by the lecturer on the whiteboard is important for indexing the scene database. We are considering the handwritten keyword recognition. The whiteboard image captured by the pen capture tool is recognized to character strings and the string corresponds to keywords in a textbook to link to the explanation of the keyword in textbook. One of important problems in our learning system is that the accuracy of handwritten character recognition on whiteboard is not enough for keyword recognition. In this paper, we propose the new matching method of high accuracy keyword recognition using word dictionary and the distance of character recognition. We confirmed the usefulness using word dictionary for handwritten keyword recognition on whiteboard

Constraints and search efficiency in nurse scheduling problem

The nurse scheduling problem (NSP) is that of allocating shifts for nurses under various constraints. Generally, the extensive number of constraints requires a great knowledge and experience to make a nurses scheduling table. This task requires a lot of time and effort and there is demand for an automatic scheduling. This study focuses the relationship between constraints and search efficiency, and discusses the affect for search efficiency, with the way to take the constraints into GA. This paper especially focuses the constraints that must be satisfied absolutely in the coding and genetic operations, and shows that his method can reduce the search area of GA drastically using numerical approximation method. Scheduling tables are acquired by the conventional and proposed method, and the result shows that effective search can be done and effective nurse scheduling tables can be generated by the proposed method.

A Study on Document Structure Recognition of Discharge Summaries for Analogous Case Search System

Recently, many paper-based documents used in hospitals are computerized because of diffusion of Clinical Information Systems (CIS). The demands for converting these documents into XML documents have been growing to re-use the information in these documents on the CIS. Previously, we proposed a recognition and XML generation method for tabular form discharge summary documents and conducted experiments to validate the effectiveness of the proposed method. The results show that documents with only characters and ruled lines are converted to XML documents appropriately. However, the documents with figures and dotted lines could not be converted in the experiment. In this paper, we propose the discrimination method to extract dotted lines and figures like schemas and so on. This paper describes the detail of the proposed method and shows the results of the experiments to validate the effectiveness of the proposed method. The problem of the proposed method and future works of this study are described at the end of this paper.

Keyword Extraction from Incident Reports and Keyword Map Generation Method Using Self Organizing Map

Recently, many hospital support systems such as electrical medical record have been computerized using information technology. In near future, all of medical data will be stored to database server, and complete paper-less and film-less system in hospitals will become reality. These stored data, however, is not used effectively for analysis because of the great deal of data and its variety - image, numerical, text data and so on. Previously, knowledge discovery methods from numerical medical database were studied widely. Text mining method for medical data, however, has not been enough discussed. This study discusses knowledge discovery method from electronic medical records using SOM and its possibility. As the first step of this research, this paper describes the outline of keyword extraction method using morphological analysis and generation of keyword map from incident report