• The Journal of the Acoustical Society of Korea
• /
• v.38 no.1
• /
• pp.39-46
• /
• 2019

In this paper, combining features is proposed as a way to enhance the classification accuracy of sounds under noisy environments using the CNN (Convolutional Neural Network) structure. A robust log Mel-filter bank using Wiener filter and PNCCs (Power Normalized Cepstral Coefficients) are extracted to form a 2-dimensional feature that is used as input to the CNN structure. An ebird database is used to classify 43 types of bird species in their natural environment. To evaluate the performance of the combined features under noisy environments, the database is augmented with 3 types of noise under 4 different SNRs (Signal to Noise Ratios) (20 dB, 10 dB, 5 dB, 0 dB). The combined feature is compared to the log Mel-filter bank with and without incorporating the Wiener filter and the PNCCs. The combined feature is shown to outperform the other mentioned features under clean environments with a 1.34 % increase in overall average accuracy. Additionally, the accuracy under noisy environments at the 4 SNR levels is increased by 1.06 % and 0.65 % for shop and schoolyard noise backgrounds, respectively.

• Journal of the Korea Society of Computer and Information
• /
• v.28 no.4
• /
• pp.187-195
• /
• 2023

This study propose to develop artificial intelligence liberal arts courses for college students in the humanities and social sciences majors using the entry artificial intelligence model. A group of experts in computer, artificial intelligence, and pedagogy was formed, and the final artificial intelligence liberal arts course was developed using previous research analysis and Delphi techniques. As a result of the study, the educational topics were largely composed of four categories: image classification, image recognition, text classification, and sound classification. The training consisted of 1) Understanding the principles of artificial intelligence, 2) Practice using the entry artificial intelligence model, 3) Identifying the Ethical Impact, and 4) Based on learned, team idea meeting to solve real-life problems. Through this course, understanding the principles of the core technology of artificial intelligence can be directly implemented through the entry artificial intelligence model, and furthermore, based on the experience of solving various real-life problems with artificial intelligence, and it can be expected to contribute positively to understanding technology, exploring the ethics needed in the artificial intelligence era.

• Phonetics and Speech Sciences
• /
• v.5 no.1
• /
• pp.55-62
• /
• 2013

This study compared the aerodynamic multiparameters of whispers and phonation in patients with muscle misuse dysphonia(MMD) to evaluate the voice aerodynamic analysis for discrimination between whispers and phonation. Eleven patients with muscle misuse dysphonia were examined. Whispers were shorter with a maximum phonation time(MPT; p<.01), a lower phonatory sound pressure level(SPLp; p<.01), a higher phonatory flow rate (PFR; p<01), lower phonatory efficiency(PE; p<.01), and a lower phonatory resistance (PR; p<.05) than phonation. The subglottal pressure level was not significantly different between whispers and phonation. (Psub; p>.05). The ROC analysis showed that the threshold of 23.83 ppm for PE achieved a good classification for whispers, with the perfect sensitivity(100%) and specificity(100%). Those results indicate PE reliably distinguished between whispers and phonation. The results also suggest that PE may provide a useful tool for studying the laryngeal source.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.06a
• /
• pp.31-32
• /
• 2017

본 논문에서는 심층 신경망 기반의 사운드 분류를 위한 청각 특성 추출 기술을 제안한다. 심층 신경망은 인간의 신경망을 모델링 하기 때문에 인간의 인식을 기반으로 하는 특성을 사용한다면 더 적합한 학습을 할 수 있다. 기존 방법인 MFCC와 스펙트로그램과는 달리 스파이크그램은 인간의 청각 시스템을 기반으로 파형을 해석하는 방법이기 때문에 심층 신경망에 더 효율적인 특성이라고 할 수 있다. 따라서 본 논문에서는 사운드 분류 기술의 특성으로 스파이크그램을 이용하는 방법을 제안한다. 제안한 방법을 사용하면 MFCC와 스펙트로그램을 사용하는 것보다 더 높은 분류 성능을 얻을 수 있다.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2006.11a
• /
• pp.757-762
• /
• 2006

The method of promoing sound surroundings with the new concept named 'Soundscape' attracts the public attention. This new method is the way of active promoting acoustic surroundings to offer the vitality, comfort and identity of urban space and relative quietness by masking the noise with natural and signal sounds. Therefore, this research has the purpose of finding out the method of soundscape concept of urban public space for the comfort acoustic surroundings.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2002.11b
• /
• pp.699-704
• /
• 2002

A method using beamforming algorithm has been developed to measure oblique incidence reflection coefficients of sound absorption materials. MUSIC(Multiple Signal Classification) method detects the angles of incidence and reflection. By separating the incident and reflected waves using beamforming method, the reflection coefficient is calculated. Spatial smoothing technique is also used to reduce the coherence between the incident and reflected waves. The test materials were modeled as a locally reacting surface. Numerical and experiment results are performed to verify the acuracy of proposed method.

• Proceedings of the KOSOMBE Conference
• /
• v.1997 no.05
• /
• pp.402-405
• /
• 1997

The analysis of heart sounds is a noninvasive diagnostic method useful to diagnose heart valve function. In this paper we compared the ability of spectral analysis method for prosthetic heart valve sounds. Phonocardiograms of prosthetic heart valve were analyzed in order to derive frequency domain feature suitable for the classification of the valve state. The FFT-based methods did not provide sufficient frequency resolution to completely characterize the spectrum of prosthetic heart valve sounds. A high resolution parametric methods were shown to give superior frequency resolution. In parametric methods, all methods provide a 1st & 2nd & 3rd frequency component. But Shank method provided a most dominant frequency peak.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.186-188
• /
• 2005

Hand region detection in images is an important process in many computer vision applications. It is a process that usually starts at a pixel-level, and that involves a pre-process of color space transformation followed by a classification process. A color space transformation is assumed to increase separability between skin classes for hands and non-skin classes for other parts, to increase similarity among different skin tones, and to bring a robust performance under varying illumination conditions, without any sound reasonings. In this work, we examine if the color space transformation does bring those benefits to the problem of hand region detection on a dataset of images with different hand postures, backgrounds, people, and illuminations. Results indicate that best of the color space is the normalized RGB.

• The Journal of the Acoustical Society of Korea
• /
• v.5 no.3
• /
• pp.28-35
• /
• 1986

우리나라 음성의 유성음을 모음, 비음 및 유성화 자음으로 분류하는 알고리즘을 기술하였다. 먼 저 기존의 PITCH 검출 알고리즘에 의하여 음성을 유성음과 무성음으로 나눈 뒤, 단지 정규화된 1차 상 관계수, 영교차율, LOG 에너지 및 LPG 에너지의 골짜기 검출만을 이용하여, 유성음은 모음, 비음 및 유 성화자음으로 분류하고 무성음은 실제의 무성음과 묵음으로 분류하였다. 그리고 이렇게 분류된 모음에 대하여 단모음 인식을 행하였다. 단지 한 FRAME으로 모음을 대표하였기 때문에 메모리 크기와 인식 시간을 줄였다. 여기서 UP & DOWN 및 수정된 영교차율을 새로이 정의하여 적용한 결과 만족한 결과 를 얻을 수 있었다. LPC 매개변수 및 전력 스펙트럼도 단모음 인식의 FEATURE로 사용하였다. 그리고 각 FEATURE 의 성능을 비교하였다. 이들 FEATURE을 잘 조합하여 2단계 인식을 행한 결과 92%의 높은 인식율을 얻을 수 있었다.

• Korean Chemical Engineering Research
• /
• v.55 no.4
• /
• pp.497-502
• /
• 2017

Estimation of accident cost is a sound and great safety indicator on determining accurate occupational safety and health prevention. Just like in Korea, Heinrich ratio analysis of (1:4) between direct and indirect costs has been become widely used in safety management because of its simplicity. In this study four major categories of uninsured (indirect) cost items and 18 sub-categories of uninsured (indirect) cost items were identified. To determine and validate the importance and necessity of the results of a literature review an expert or professional surveyed had been analyses using the SPSS 18.0, where in the participants whose expertize is in the field of compensation and safety. Based on the results of survey all participants all uninsured (indirect) cost items classified was important and necessary when accidents occurred. Despite recognition of expert on the classification of uninsured (indirect) cost items, it is quite difficult to make generalization for all kind of costs in occupational accident case due to different nature of business for each industry.