• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.559-561
• /
• 2022

인공지능 시대에 들어서면서 개인 맞춤형 환경을 제공하기 위하여 사람의 감정을 인식하고 교감하는 기술이 많이 발전되고 있다. 사람의 감정을 인식하는 방법으로는 얼굴, 음성, 신체 동작, 생체 신호 등이 있지만 이 중 가장 직관적이면서도 쉽게 접할 수 있는 것은 표정이다. 따라서, 본 논문에서는 정확도 높은 얼굴 감정 식별을 위해서 Convolution Block Attention Module(CBAM)의 각 Gate와 Residual Block, Skip Connection을 이용한 Identity- CBAM Module을 제안한다. CBAM의 각 Gate와 Residual Block을 이용하여 각각의 표정에 대한 핵심 특징 정보들을 강조하여 Context 한 모델로 변화시켜주는 효과를 가지게 하였으며 Skip-Connection을 이용하여 기울기 소실 및 폭발에 강인하게 해주는 모듈을 제안한다. AI-HUB의 한국인 감정 인식을 위한 복합 영상 데이터 세트를 이용하여 총 6개의 클래스로 구분하였으며, F1-Score, Accuracy 기준으로 Identity-CBAM 모듈을 적용하였을 때 Vanilla ResNet50, ResNet101 대비 F1-Score 0.4~2.7%, Accuracy 0.18~2.03%의 성능 향상을 달성하였다. 또한, Guided Backpropagation과 Guided GradCam을 통해 시각화하였을 때 중요 특징점들을 더 세밀하게 표현하는 것을 확인하였다. 결과적으로 이미지 내 표정 분류 Task에서 Vanilla ResNet50, ResNet101을 사용하는 것보다 Identity-CBAM Module을 함께 사용하는 것이 더 적합함을 입증하였다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.5
• /
• pp.963-969
• /
• 2000

In this paper, we define two input-output gains of linear periodic time-varying systems. One is the ratio of output with worst-case l2-norm over all inputs with unit 12-norm. It denotes G($\iota_2,\iota_2$.The other is the ratio of output with worst-case RMS value over all inputs with unit RMS value. It denotes G(RMS, RMS) .It is fact that these two gains are equivalent for linear time-invariant system. In this paper, we prove these two gains are also equivalent for linear periodic time-varying system. In addition, the relationship between two method of obtaining the generalized frequency responses for linear periodic time-varying system is derived. Finally, we apply the defined input-output gains to M-channel filter-bank which is multi-rate signal Processing system, used to speech coding. In the filter-bank, generally, aliasing distortion, magnitude distortion, and phase distortion are present. It is shown that these are kept small if the filter-bank is designed by a method that optimizes the gain G($\iota_2,\iota_2$ of an error system.

• Journal of Korea Multimedia Society
• /
• v.3 no.4
• /
• pp.347-355
• /
• 2000

This paper presents an adaptive AEC(acoustic echo canceller) based on the wavelet transform using M-channel subband QMF filter banks. The proposed algorithm improves the performance of AEC with a realtime process by a low complexity of wavelet transform filter banks, a subband processing and a orthogonality of wavelet subband filter. Adaptive filter coefficients of each subband are updated using LMS algorithm with a low complexity and a easy realization for a realtime processing and a reduction of hardware cost. For a input signal, a white Gaussian noise and a real speech signal with a environment noises are used for a performance estimation of the proposed algorithm. As a result of computer simulation, the proposed AEC has a low asymptotic error, a low computation complexity and a robust performance.

• Journal of the Institute of Convergence Signal Processing
• /
• v.2 no.2
• /
• pp.62-67
• /
• 2001

In the recent mobile communication systems, the performance of Turbo Code using the error correction coding depends on the interleaver influencing the free distance determination and the recursive decoding algorithms that is executed in the turbo decoder. However, performance depends on the interleaver depth that need a large time delay over the reception process. Moreover, Turbo Code has been known as the robust ending method with the confidence over the fading channel. The International Telecommunication Union(ITU) has recently adopted as the standardization of the channel coding over the third generation mobile communications such as IMT-2000. Therefore, in this paper, we proposed of the method to improve the conventional performance with the parallel concatenated 4-New Turbo Decoder using MAP a1gorithm in spite of complexity increasement. In the real-time video and video service over the third generation mobile communications, the performance of the proposed method was analyzed by the reduced decoding delay using the variable decoding method by computer simulation over AWGN and fading channels.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.12a
• /
• pp.13-16
• /
• 2000

In the recent mobile communication systems, the performance of Turbo Code using the error correction coding depends on the interleaver influencing the free distance determination and the recursive decoding algorithms that is executed in the turbo decoder. However, performance depends on the interleaver depth that need great many delay over the reception process. Moreover, Turbo Code has been known as the robust coding methods with the confidence over the fading channel. The International Telecommunication Union(ITU) has recently adopted as the standardization of the channel coding over the third generation mobile communications the same as IMT-2000. Therefore, in this paper, we proposed of that has the better performance than existing Turbo Decoder that has the parallel concatenated four-step structure using MAP algorithm. In the real-time voice and video service over the third generation mobile communications, the performance of the proposed method was analyzed by the reduced decoding delay using the variable decoding method by computer simulation over AWGN and lading channels.

• Journal of the Institute of Convergence Signal Processing
• /
• v.22 no.2
• /
• pp.57-63
• /
• 2021

The purpose of this study is to develop a Smart Helmet to safeguard forest firefighting crews and provide on-site information in real time. The Smart Helmet for forest firefingting crews is equipped with a camera, video/voice communication module, GPS, Bluetooth, and LTE module to promote the safety of them, and through the Smart Helmet, the site situation is is transmitted in real time, and full duplex communication is possible. As a result of testing using the Smart Helmet, the control center was able to receive on-site information and communication with on-site forest firefighting crews. Through site evaluation and user evaluation, it was confirmed that the Smart Helmet needs to be improved. The developed Smart Helmet can be used in various ways in forest disasters and forest industry.

• Journal of the Institute of Convergence Signal Processing
• /
• v.24 no.2
• /
• pp.82-89
• /
• 2023

Due to the characteristics of ship structure, the compartment structure is complicated and narrow, so safety accidents frequently occur during the work process. The main causes of accidents include structural collisions, falling objects, toxic substance leaks, fires, explosions, asphyxiation, and more. Understanding the on-site conditions of workers during accidents is crucial for mitigating damages. In order to ensure safety, the on-site situation is monitored using CCTV in the ship, but it is difficult to prevent accidents with the existing method. To address this issue, a smart safety helmet equipped with location identification and voice/video communication capabilities is being developed as a safety technology. Additionally, the smart safety helmet incorporates environmental sensors for temperature, humidity, vibration, noise, tilt (gyro sensor), and gas detection within the work area. These sensors can notify workers wearing the smart safety helmet of hazardous situations. By utilizing the smart safety helmet and environmental sensors, the safety of workers aboard ships can be enhanced.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.8
• /
• pp.260-267
• /
• 2018

Surface electromyography (sEMG) is mainly used as an on/off switch in the early stage of the study and was then expanded to navigational control of powered-wheelchairs and recognition of sign language or finger gestures. There are difficulties in communication between people who know and do not know sign language; therefore, many efforts have been made to recognize sign language or finger gestures. Recently, use of sEMG signals to recognize sign language signals have been investigated; however, most studies of this topic conducted to date have focused on Chinese finger number gestures. Since sign language and finger gestures vary among regions, Korean- and Chinese-finger number gestures differ from each other. Accordingly, the recognition performance of Korean finger number gestures based on sEMG signals can be severely degraded if the same muscles are specified as for Chinese finger number gestures. However, few studies of Korean finger number gestures based on sEMG signals have been conducted. Thus, this study was conducted to identify potential forearm muscles from which to collect sEMG signals for Korean finger number gestures. To accomplish this, six Korean finger number gestures from number zero to five were investigated to determine the usefulness of the proposed muscles and electrode placements by showing that CNN technique based on sEMG signal after sufficient learning recognizes six Korean finger number gestures in accuracy of 100%.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.7
• /
• pp.1575-1581
• /
• 2003

As a society has progressed rapidly toward a highly advanced digital information age, a multimedia communication service for acquisition, transmission and storage of image data as well as voice has being commercialized externally and internally. However, in the process of digitalization or transmission of data, noise is generated by several causes, and researches for eliminating those noises have been continued until now. There were the existing FFT(fast fourier transform) and STFT(short time fourier transform) for removing noise but it's impossible to know information about time and time-frequency localization capabilities has conflictive relationship. Therefore, for overcoming these limits, wavelet transform which is presented as a new technique of signal processing field is being applied in many fields recently. Because it has time-frequency localization capabilities it's Possible for multiresolution analysis as well as easy to analyze various signal. And when two wavelet base were designed to form Hilbert transform pair, wavelet pair provide superior performance than the existing DWT(discrete wavelet transform) in data characteristic detection. Therefore in this parer, we removed impulse noise by using adaptive-length median filter and two dyadic wavelet base which is designed by truncated coefficient vector.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.1 no.1
• /
• pp.70-80
• /
• 2006

The large portion of communication service areas move from a legacy wire-line voice service to mobile data service. For the purpose of satisfaction on market need, many communication systems should be installed and upgraded based on a mobile wide-band transmission facility. Recently, large part of communication service is based on internet protocol by packet switch techniques and required new technologies such as multimedia processing, QoS achievement, and mobility managememobile communication network such as IS-95A/B and CDMA2000-1X. In this paper, I analyzed the network architecture and service provision methods. in CDMA2000-1X nt. In korea, a CDMA communication technique is standardized for digital mobile communication systems. By using the analysed results, I will extract an efficient method for network evolution and a core technique for next generation mobile communication network.