Browsing by Author "Jong, Tai-Lang"

Now showing 1 - 4 of 4

The prompt of lip shape modification of cacology based on the speech evaluation techniques - a case of basic Chinese learning
(2008)
Hsieh, Chi-Wen
;
Lin, Chih-Huang
;
Jong, Tai-Lang
;
Hsieh, Chi-Yi
;
電子物理學系
In the study, a Chinese learning assisted system based on speech recognition and lip shape image processing is proposed. The mel-frequency cepstral coefficient (MFCC), the pitch contour, and energy curve were adopted as the parameters of voiceprint, speech tone, and magnitude of speech signals, respectively. On the other hand, the height and width of the lip shape were sent into the lip shape analysis. In the scoring stage of speech utterances, the dynamic time warping (DTW) algorithm and probabilistic neural network (PNN) were applied to determine whether the test speech was qualified or not during Chinese learning process. The simulation results indicated that the hybrid of MFCC, pitch contour, and energy curve parameters of speech signal could slightly promote the accuracy, of classification-could achieve up to 90%. Finally, the Receiver Operating Characteristic Curve (ROC) was introduced to quantitatively evaluate the sensitivity and specificity of the performance of the proposed algorithm.
The study of the relationship between electro-oculogram and the features of closed eye motion
(2008)
Hsieh, Chi-Wen
;
Chen, Hsiaw-Shuw
;
Jong, Tai-Lang
;
電子物理學系
;
Department of Electrophysics
The study emphasized on the possibility of using computer vision to synthesize the electro-oculograrn (EOG) signals acquired from the contact electrode measuring. The synchronized image frames of the closed eye status and the corresponding EOG signals were sent into the re-sampling module to be aligned. Four categories of images features including spatial domain, frequency domain, statistics, and local entropy analysis were adopted to transfer the two-dimensional image into a time sequential signal and to make the comparison between EOG signals and corresponding features. The correlation analysis was recruited to qualitatively identify the relationship of EOG signals and closed eyeball moving images. The simulation results indicates that the local entropy feature has a close correspondence with EOG signal; as a result, the proposed fuzzy membership function based on the local entropy blocks might deliver a non-invasively synthesized EOG signal in sleep research practically.
Turner syndrome phalangeal screening based on a two-stage linear regression concept
(2009-08-01)
Tiu, Chui-Mei
;
Liu, Tzu-Chiang
;
Hsieh, Chi-Wen
;
Niu, Dau-Ming
;
Chen, Jen-Dar
Background: Turner syndrome (TS) is a congenital chromosomal abnormality, resulting in short stature, short fourth metacarpal, and retarded skeletal maturation in children. The existing methods of diagnosis, which include carpal angle, metacarpal sign, and body mass index (BMI), cannot accurately diagnose TS. The authors propose a novel procedure for examining the hand skeleton to distinguish between normal individuals and patients with TS. Methods: This investigation was divided into two parts. In the first part, existing methods (evaluation of the metacarpal sign, measurement of the carpal angle, and determination of BMI) were used. Examination in the second part was based on the two-stage screening method (TSSM). In the first stage in TSSM, the ratio of the lengths of the distal-middle phalanges of the fifth digit was determined in normal subjects with average range of satisfactory body height and TS patients. A suitable cut-off was found on linear regression and used to divide the plot into TS patients and normal subjects. In the second stage, the normal section was transferred to another group based on bone and chronological ages. A greater number of patients were diagnosed with TS using this method. Finally, four cut-off parameters were determined on linear regression analysis. Results with optimal sensitivity and specificity were automatically obtained. Results: The combination of TSSM with optimal programming (sensitivity = 0.81 and specificity = 0.91) was satisfactory for diagnosing TS patients. Conclusion: TSSM can suitably evaluate growth of the hand skeleton to distinguish between normal individuals and patients with TS.
Turner's syndrome screening based on hand image assessment
(2008)
Hsieh, Chi-Wen
;
Liu, Tzu-Chiang
;
Jong, Tai-Lang
;
Tiu, Chui-Mei
;
電子物理學系
The study presents a algorithm to screen the children with Turner's Syndrome based on the hand digital images analysis. The ratio of the distal-middle phalanx of the left fifth finger versus chronological age in normal children and children with Turner's syndrome were analyzed. Most Turner syndrome cases mixed with few normal cases are identifed in the first screening process based on phalanx's length analysis. The relationship between bone age and chronological age is also studied to locate the TS patterns precisely. Linear regression analysis is used to find the boundary lines to promote the accuracy in two screening stages. The preliminary result indicates the metacarpal sign, carpal angle, and body mass index are not significant parameters to the TS screening. On the other hand, the sensitivity and specificity can achieve to 0.81 and 0.91, respectively.