• 음성과학
• /
• 제4권2호
• /
• pp.17-24
• /
• 1998

In this paper a method to diagnose pathological voices using wavelet transform is sug gested. Pathological voices are collected from hospital and analyzed by the suggested method. Normal voices are collected separately and analyzed. Then the results are compared to find the differences in their characteristics. Three level wavelet transform is used. Normalized energy ratios between the levels and normalized peak-to-peak values are used as parameters. As a result, it was possible to distinguish between normal and pathological voices.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 제38회 하계학술대회
• /
• pp.1797-1798
• /
• 2007

In field of voice pathology, diverse statistics extracted form pitch estimation were commonly used to assess voice quality. In this study, we proposed robust pitch detection algorithm which can estimate pitch of pathological voices in benign vocal fold lesions. we also compared our proposed algorithm with three established pitch detection algorithms; autocorrelation, simplified inverse filtering technique, and nonlinear state-space embedding methods. 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. 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.

• 말소리와 음성과학
• /
• 제2권1호
• /
• pp.113-120
• /
• 2010

In this paper, we investigate the acoustic characteristics of sustained voices from normal subjects and patients with laryngeal pathologies. Perturbation methods (including jitter and shimmer), signal-to-noise ratio (SNR), and nonlinear dynamic methods (such as correlation dimension) are used to analyze normal and pathological voices. We find that jitter does not statistically discriminate between normal and pathological voices, but a significant difference is found for shimmer, SNR, and correlation dimension. The results suggest that nonlinear dynamic analysis may be valuable for the analysis of normal and pathological voices but perturbation analysis should be applied with caution for pathological voice analysis.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2007년도 하계종합학술대회 논문집
• /
• pp.407-408
• /
• 2007

Robust pitch estimation is an important study in many areas of speech processing. In voice pathology, diverse statistics extracted form pitch 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 such as benign vocal fold lesions; polyp, nodule, and cysts. Consequently, it is required to survey the severity of tested voice in order to obtain accurate pitch estimates.

• 음성과학
• /
• 제14권4호
• /
• pp.179-200
• /
• 2007

In voice pathology, various measurements calculated from pitch values are proposed to show voice quality. However, those measurements frequently seem to be inaccurate and unreliable because they are based on some wrong pitch values determined from pathological voice data. In order to solve the problem, we compared several pitch estimation methods to propose a better one in pathological voices. From the database of 99 pathological voice and 30 normal voice data, errors derived from pitch estimation were analyzed and compared between pathological and normal voice data or among the vowels produced by patients with benign vocal fold lesions. Results showed that gross pitch errors were observed in the cases of pathological voice data. From the types of pathological voices classified by the degree of aperiodicity in the speech signals, we found that pitch errors were closely related to the number of aperiodic segments. Also, the autocorrelation approach was found to be the most robust pitch estimation in the pathological voice data. It is desirable to conduct further research on the more severely pathological voice data in order to reduce pitch estimation errors.

• 말소리와 음성과학
• /
• 제2권2호
• /
• pp.109-115
• /
• 2010

In this paper, we investigate a signal typing on the basis of visual impression of distinctive spectrogram. Pathological voices are classified into signal type 1, 2, 3, or 4 to estimate perturbation parameters and to mark perceptual rating based on Consensus Auditory-Perceptual Evaluation of Voice (CAPE-V). The results suggest that perturbation analysis can be applied to only type 1 and 2 signals and the perceptual ratings of overall grade increase with each signal type, overall. A good inter-rater reliability is showed among three raters. We recommend that pathological voices should be marked the signal typing and CAPE-V, together, to definitely describe the characteristics of pathological voices.

• 말소리와 음성과학
• /
• 제3권3호
• /
• pp.99-108
• /
• 2011

The perturbation parameters like jitter, shimmer, and signal-to-noise ratio (SNR) are largely estimated in the particular segment from the subjective or whole portion of the given pathological voice signal although there are many possible regions to be able to analyze the voice signals. In this paper, the pathological voice signals were classified as type 1, 2, 3, or 4 according to narrow band spectrogram and the value differences of the perturbation parameters extracted in the subjective and entire portion tended to be getting bigger as from type 1 to type 4 signals. Therefore, sample selection method based on moving window to analyze type 2 and 3 signals as well as type 1 signals is proposed. Although type 3 signals cannot be analyzed using the perturbation analysis, the type 3 signals by selecting out the samples in which error count is less than 10 through moving window were analyzed. At present, there is no method to be able to analyze the type 4 signals. Future research will endeavor to determine the best way to evaluate such voices.

• 대한후두음성언어의학회지
• /
• 제33권3호
• /
• pp.142-155
• /
• 2022

Diagnosis using voice is non-invasive and can be implemented through various voice recording devices; therefore, it can be used as a screening or diagnostic assistant tool for laryngeal voice disease to help clinicians. The development of artificial intelligence algorithms, such as machine learning, led by the latest deep learning technology, began with a binary classification that distinguishes normal and pathological voices; consequently, it has contributed in improving the accuracy of multi-classification to classify various types of pathological voices. However, no conclusions that can be applied in the clinical field have yet been achieved. Most studies on pathological speech classification using speech have used the continuous short vowel /ah/, which is relatively easier than using continuous or running speech. However, continuous speech has the potential to derive more accurate results as additional information can be obtained from the change in the voice signal over time. In this review, explanations of terms related to artificial intelligence research, and the latest trends in machine learning and deep learning algorithms are reviewed; furthermore, the latest research results and limitations are introduced to provide future directions for researchers.

• 디지털융복합연구
• /
• 제13권12호
• /
• pp.187-192
• /
• 2015

장애음성을 판별할 수 있는 소프트웨어가 개발 될 경우, 원격의료와 언어치료 등 여러 융복합 분야에서의 활용도가 매우 높다. 본 논문은 성대 진동에 대한 변화율을 나타내는 의료정보인 음향학적 파라미터와 신호처리 기반 고차 통계량에 기반을 둔 파라미터를 융합하여, CART(Classification And Regression Trees) 분석을 통해서 정상/장애음성 판별 프로그램을 구현하였다. 사용된 음향학적 파라미터는 Jitter(%)와 shimmer(%)이다. 그리고 본 연구에서 제안된 고차통계량 기반 파라미터는 왜도(Skewness)와 첨도(Kurtosis)의 평균과 분산이다. Kay Elemetrics의 데이터베이스에서 무작위로 발췌된 정상음성 53명, 장애 음성 173명의 /아/ 발화를 이용하여 결정트리(Decision tree) 기반장애음성 판별을 위해 평균적으로 83.15%의 성능을 보이는 알고리즘을 구현하였다. 그 결과를 바탕으로 추후 상용화를 고려하여 사용자 친화적인 프레임 워크에 의해 컨텐츠를 생성하는 융복합형 기능이 포함된 장애음성 판별 프로그램을 개발하였다.

• 대한후두음성언어의학회지
• /
• 제15권2호
• /
• pp.92-97
• /
• 2004

Background and Objectives : The purpose of this study was to examine objectively pre and post operative voice quality evaluation and intelligibility of alaryngeal voice using speech recognition program, HUVOIS. Materials and Methods : 2 laryngologists and 1 speech pathologist were evaluated 'G', 'R', 'B' in the GRBAS sclae and speech intelligibility using NTID rating scale from standard paragraph. And also acoustic estimates such as jitter, shimmer, HNR were obtained from Lx Speech Studio. Results : Speech recognition rate was not significantly different between pre and post operation for pathological vocie samples though voice quality(G, B) and acoustic values(Jitter, HNR) were significantly improved after post operation. In Alaryngeal voices, reed type electrolarynx 'Moksori' was the highest both speech intelligibility and speech recognition rate, whereas esophageal speech was the lowest. Coefficient correlation of speech intelligibility and speech recognition rate was found in alaryngeal voices, but not in pathological voices. Conclusion : Current study was not proved speech recognition program, HUVOIS during telephone program was not objective and efficient method for assisting subjective GRBAS scale.