Seung-Jin Jang

Evaluation of Performance of Several Established Pitch Detection Algorithms in Pathological Voices

Robust pitch estimation is important in many areas of speech processing. In voice pathology, diverse statistics extracted form pitch estimation were commonly used to test voice quality. In this study, we compared several established pitch detection algorithms (PDAs) for verification of adequacy of the PDAs. In the database of total pathological voices of 99 and normal voices of 30, an analysis of errors related with pitch detection was evaluated between pathological and normal voices, or among the types of pathological voices. Pitch errors of all PDAs used in this study more or less showed some changes between pathological and normal voices. According to the results of pitch errors, gross pitch error showed some increases in cases of pathological voices; especially excessive increase in PDA based on nonlinear time-series. In an analysis of types of pathological voices classified by aperiodicity and the degree of chaos, the more voice has aperiodic and chaotic, the more growth of pitch errors increased. Consequently, it is required to survey the severity of tested voice in order to obtain accurate pitch estimates.

Sustained Vowel Modeling using Nonlinear Autoregressive Method based on Least Squares-Support Vector Regression

In this paper, Nonlinear Autoregressive (NAR) method based on Least Square-Support Vector Regression (LS-SVR) is introduced and tested for nonlinear sustained vowel modeling. In the database of total 43 sustained vowel of Benign Vocal Fold Lesions having aperiodic waveform, this nonlinear synthesizer near perfectly reproduced chaotic sustained vowels, and also conserved the naturalness of sound such as jitter, compared to Linear Predictive Coding does not keep these naturalness. However, the results of some phonation are quite different from the original sounds. These results are assumed that single-band model can not afford to control and decompose the high frequency components. Therefore multi-band model with wavelet filterbank is adopted for substituting single band model. As a results, multi-band model results in improved stability. Finally, nonlinear sustained vowel modeling using NAR based on LS-SVR can successfully reconstruct synthesized sounds nearly similar to original voiced sounds.

Upiquitous Home Healthcare Management System with Early Warning Reporting

The changes of medical environments such as the growth of a patient with a chronic disease and the elderly, and the advance of the IT technology require highly qualified medical service of new paradigm. For the reason, home health management service has been focused with much expectations and interests. Our proposed home health management system (H2MS) monitors and manages daily physiological status which is easily acquired from multiple-sensor based ubiquitous bio-signal sensor during sleeping, bathing, evacuating, and living at home. H2MS is largely consisted of four functional subsystems: multiple-sensor based ubiquitous bio-signal sensor, home health intermediary client (H2IC), home health center (H2C), and home health service provider (H2SP). Each subsystem carried out their own processes and collaborated on a work with each other. H2MS provides various home health services for the wellness, as well as the elderly, chronic patients illness, or terminal illness. Also, it is available to suggest various services: an early warning service according to the change of health status, advice service for health improvement, and appointment service for medical treatment. Moreover, this early warning service makes promptly emergency repairs as soon as occurrence of physiologically abnormal symptom. Consequently, early warning reports reduce the frequent intervention of high costed service of medical professionals.

Analysis of relation between coronary perfusion pressure and the extracted parameters from a ventricular fibrillation ECG signal

This work presents an alternative return of spontaneous circulation (ROSC) estimate using indirectly induced presumption that coronary perfusion pressure (CPP) correlates with the extracted parameter from the ventricular fibrillation (VF) ECG signal. In past studies, it is revealed that successful cardiopulmonary resuscitation (CPR) needs at least 30/spl sim/40mmHg CPP during the aortic diastolic period. In 360 segments derived from 18 test dogs with experimental cardiac arrest of cardiac cause, we analyzed the ability of 4 spectral features of VF before countershock to discriminate or not between segments that correspond to CPP. The median frequency (MF), peak frequency (PF), average segment amplitude (ASA) and maximum segment amplitude (MSA) were studied. After preprocessing the raw data acquired from the specific experimental setup and protocol, we verified CPP is a serious estimate of ROSC, and then we analyzed the extracted parameters corresponding to CPP by multiple regression. In the specific conditional frequency domain (MF: 9.42/spl sim/12.42Hz, PF: 8.71/spl sim/13.08Hz, ASA: > 0.19mV), CPP is correlated to the extracted parameter with 0.71 /spl plusmn/ 0.05 coefficient of multiple determination (R/sup 2/). The combination of MF, PF, and ASA achieved a 79.47 /spl plusmn/ 3% sensitivity and 41.67 /spl plusmn/ 4% specificity in testing.

XML based on-line Clinical Decision Support System for Home Healthcare

This paper describes the process and methodology of designing and developing a clinical decision support system for home healthcare system for multi user that classify the patients into proper health level and handle the on-line monitoring, a reaction of the classification result. H2MRC Center is home healthcare system that acquires fundamental patient’s important 8 health parameters and monitors the patient’s information and classifies befitting health level and responses in health level. H2MRC have 8 important key health information, 8 parameters that classified A. the sap of life, B. Respiratory, C. Cardiovascular, D. Endocrines, and So, it needs a fast the classification of health level result, automatic classification, reliability of the result. But many users participate in H2MRC program and 8 parameters each have complicated relations so hard to monitor by a physician. So we design XML based SDSS for home healthcare. It helps fast classified by health level using these acquired 8 parameters, and suggests the proper diagnosis and provides proper medical information that helps patient’s Diagnosis. Also it helps the physician’s effectiveness of on-line decision support using large volumes of data that has complicated relation.

Design and Implementation of a Lavatory Information System Using RFID Technology

Laboratory information System (LIS) is a key tool for management of laboratory data and for improvement of service quality in clinical pathology. But various type, complexity, and large quantity of specimen, occur missing the specimen or report. So we suggest the RFID based LIS system. We developed implementation LIS system has better CQI(Continuous Quality Improvement) using RFID and XML, ADO.net, ASP.net c# 2.0. RFID technology replaces to an existing the bar-code system. Also for more convenience, we used PDA RFID module. The LIS system consists of Main Server, SQL 2005, event display and analysis web page developed using asp.net 2.0 and sharing specimen data information using XML technology.