• The Journal of the Acoustical Society of Korea
• /
• 제36권4호
• /
• pp.267-272
• /
• 2017

In this paper, we propose an effective method of applying multichannel-audio feature values to GRNNs (Gated Recurrent Neural Networks) in polyphonic sound event detection. Real life sounds are often overlapped with each other, so that it is difficult to distinguish them by using a mono-channel audio features. In the proposed method, we tried to improve the performance of polyphonic sound event detection by using multi-channel audio features. In addition, we also tried to improve the performance of polyphonic sound event detection by applying a gated recurrent neural network which is simpler than LSTM (Long Short Term Memory), which shows the highest performance among the current recurrent neural networks. The experimental results show that the proposed method achieves better sound event detection performance than other existing methods.

• The Journal of the Acoustical Society of Korea
• /
• 제39권6호
• /
• pp.600-605
• /
• 2020

In this paper, we propose a sound event detection method using a multi-channel multi-scale neural networks for sound sensing home monitoring for the hearing impaired. In the proposed system, two channels with high signal quality are selected from several wireless microphone sensors in home. The three features (time difference of arrival, pitch range, and outputs obtained by applying multi-scale convolutional neural network to log mel spectrogram) extracted from the sensor signals are applied to a classifier based on a bidirectional gated recurrent neural network to further improve the performance of sound event detection. The detected sound event result is converted into text along with the sensor position of the selected channel and provided to the hearing impaired. The experimental results show that the sound event detection method of the proposed system is superior to the existing method and can effectively deliver sound information to the hearing impaired.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 한국해양정보통신학회 2008년도 춘계종합학술대회 A
• /
• pp.805-808
• /
• 2008

In this paper, it was performed the analysis on various intersection acoustic patterns for detection rate improvement of accident sound detection system : an acoustic pattern analysis on general traffic noise, an acoustic pattern analysis on engine noise, an acoustic pattern analysis on obstruct factors for accident sound detection system. There are remarkable differences between the acoustic patterns of traffic noise and accident sound, and we most consider the acoustic patterns when we compose the accident traffic detection system by acoustic because there is error range of 20[dB] according to the volume of traffic in intersection.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• 제49권4호
• /
• pp.220-224
• /
• 2000

A new auscultation system for the detection of breath sound form trachea was developed in house. Small size microphone(panasonic pin microphone) was encapsuled in a housing for resonant effect, and hardware for the sound detection was fabricated. Pulmonary function test results were compared with the parameters extracted from frequency spectrum of breath sound obtained from the developed system. Results showed that the peak frequency and relative ratio of integral values between low(80∼400Hz) and high(400∼800Hz) frequency ranges revealed the significant differences. Developed system could be used for distinguishing normal subject and the patients who have pulmonary disease.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2009년도 학술대회
• /
• pp.158-162
• /
• 2009

Sound effects coming with appropriate tactile stimuli can strengthen its reality. For example, gunfire in games and movies, if it is accompanied by vibrating effects, can enhance the impressiveness. On a similar principle, adding the vibration information to existing sound data file and playing sound while generating vibration effects through haptic interfaces can augment the sound experience. In this paper, we propose a method to generate vibration information by analyzing the sound. The vibration information consists of vibration patterns and the timing within a sound file. Adding the vibration information is labor-intensive if it is done manually. We propose a sound detection algorithm to search the moments when specific sounds occur in a sound file and a method to create vibration effects at those moments. The sound detection algorithm compares the frequency characteristic of specific sounds and finds the moments which have similar frequency characteristic within a sound file. The detection ratio of the algorithm was 98% for five different kinds of gunfire. We also develop a GUI based vibrating pattern editor to easily perform the sound search and vibration generation.

• ETRI Journal
• /
• 제39권6호
• /
• pp.832-840
• /
• 2017

Recently, deep recurrent neural networks have achieved great success in various machine learning tasks, and have also been applied for sound event detection. The detection of temporally overlapping sound events in realistic environments is much more challenging than in monophonic detection problems. In this paper, we present an approach to improve the accuracy of polyphonic sound event detection in multichannel audio based on gated recurrent neural networks in combination with auditory spectral features. In the proposed method, human hearing perception-based spatial and spectral-domain noise-reduced harmonic features are extracted from multichannel audio and used as high-resolution spectral inputs to train gated recurrent neural networks. This provides a fast and stable convergence rate compared to long short-term memory recurrent neural networks. Our evaluation reveals that the proposed method outperforms the conventional approaches.

• Journal of Convergence for Information Technology
• /
• 제10권8호
• /
• pp.53-59
• /
• 2020

This paper relates to an IoT device system that detects sound source and estimates the sound source location. More specifically, it is a system using a sound source direction detection device that can accurately detect the direction of a sound source by analyzing the difference of arrival time of a sound source signal collected from microphone sensors, and track the generation direction of a sound source using an IoT sensor. As a result of a performance test by generating a sound source, it was confirmed that it operates very accurately within 140dB of the acoustic detection area, within 1 second of response time, and within 1° of directional angle resolution. In the future, based on this design plan, we plan to commercialize it by improving the reliability by reflecting the artificial intelligence algorithm through big data analysis.

• MALSORI
• /
• 제49호
• /
• pp.135-144
• /
• 2004

In this paper, we make a study of classification normal from abnormal - normal, asthma through analysis of thoracic sound to take use thoracic sound detection system. Thoracic sound detection system has a function to store thoracic sound and analyze the data. The wave shape of thoracic sound is similar to noise and is systematically generated by inhalation and exhalation breathing, therefore, in this paper, to classify asthma sound in thoracic sound, we could discriminate between normal and abnormal case using level crossing rate(LCR) and spectrogram energy rate.

• Journal of Advanced Marine Engineering and Technology
• /
• 제23권4호
• /
• pp.473-479
• /
• 1999

The two-microphone sound-intensity technique has been used for the detection of defects in ra-ally loaded ball bearings. The difference in the sound-intensity levels measured for bearings with no defect and for those with intentionally introduced defects of different sizes n heir elements under various operating conditions of loads and speeds is demonstrated. The results show that of an inner-race or ball defect. It is difficult to detect defects at lower speeds. Sound-pressure measurements were also performed for comparison and it shown that the detectability of defects by sound-intensity measurements is better than that by sound-pressure measurements.

• The Journal of the Acoustical Society of Korea
• /
• 제41권6호
• /
• pp.601-609
• /
• 2022

In August 2007, coastal upwelling occurred off the east coast of Korea, and vertical water temperature and salinity data were obtained from a real-time surface ocean buoy. Based on the time series observation data, a vertical sound velocity structure was calculated before, during, and after the occurrence of the coastal upwelling, and how the coastal upwelling affects the sound propagation and detection environment through acoustic modeling considering the horizontal scale and actual seabed topography. As a result of comparing and analyzing the low-frequency (500 Hz) sound transmission loss and the target detection range by depth using the parabolic equation model, it was analyzed that if coastal upwelling occurs, a detection gain of up to about 10 dB can be expected. In addition, through this study, it was confirmed that the characteristics of sound propagation can be greatly changed even in a short period of about 2 to 3 days before and after coastal upwelling.