• Journal of KIISE:Software and Applications
• /
• v.28 no.2
• /
• pp.192-200
• /
• 2001

본 논문은 조명 변화, 표정 변화, 부분적인 오클루전이 있는 얼굴 영상에 강인하고 적은 메모리양과 계산량을 갖는 효율적인 얼굴 인식 방법을 제안한다. SKKUface(Sungkyunkwan University face)라 명명한 이 방법은 먼저 훈련 영상에 PCA(principal component analysis)를 적용하여 차원을 줄일 때 구해지는 특징 벡터 공간에서 조명 변화, 얼굴 표정 변화 등에 해당되는 공간이 최대한 제외된 새로운 특징 벡터 공간을 생성한다. 이러한 특징 벡터 공간은 얼굴의 고유특징만을 주로 포함하는 벡터 공간이므로 이러한 벡터 공간에 Fisher linear discriminant를 적용하면 클래스간의 더욱 효과적인 분리가 이루어져 인식률을 획기적으로 향상시킨다. 또한, SKKUface 방법은 클래스간 분산(between-class covariance) 행렬과 클래스내 분산(within-class covariance) 행렬을 계산할 때 문제가 되는 메모리양과 계산 시간을 획기적으로 줄이는 방법을 제안하여 적용하였다. 제안된 SKKUface 방법의 얼굴 인식 성능을 평가하기 위하여 YALE, SKKU, ORL(Olivetti Research Laboratory) 얼굴 데이타베이스를 가지고 기존의 얼굴 인식 방법으로 널리 알려진 Eigenface 방법, Fisherface 방법과 함께 인식률을 비교 평가하였다. 실험 결과, 제안된 SKKUface 방법이 조명 변화, 부분적인 오클루전이 있는 얼굴 영상에 대해서 Eigenface 방법과 Fisherface 방법에 비해 인식률이 상당히 우수함을 알 수 있었다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.3
• /
• pp.175-182
• /
• 2004

There is a growing interest in speaker verification, which verifies someone by his/her voices. There are many speaker vitrification algorithms such as HMM and DTW. However, it is impossible to apply these algorithms to memory limited applications because of large number of feature vectors to register or verify users. In this paper we introduces a speaker verification system using SVM, which needs a little memory usage and computation time. Also we proposed hardware architecture for SVM. Experiments were conducted with Korean database which consists of four-digit strings. Although the error rate of SVM is slightly higher than that of HMM, SVM required much less computation time and small model size.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06e
• /
• pp.385-388
• /
• 1998

DTW를 이용한 음성인식에서는 참조패턴이 인식률에 절대적인 영향을 미치므로 가장 적합한 참조패턴의 생성이 중요한 요인으로 작용한다. 그러므로 인식률 향상을 위해 여러개의 참조패턴을 사용하는 방법이 있다. 그러나 이러한 방법은 게산량의 과다 및 사용 메모리의 증가 등이 단점으로 지적되고 있다. 따라서 본 논문에서는 참조패턴의 수를 줄이면서 높은 인식률을 얻기 위해 유전자 알고리듬을 이용하여 보다 우수한 참조패턴을 생성하여 음성인식에 적용하였다. 본 논문에서는 참조패턴 생성을 위하여 훈련에 참가한 자료를 서로 비교하여 DTW 거리값의 누적값이 최소가 되는 데이터를 선정하는 방법, 유전자 알고리듬을 이용한 방법으로 선정하는 방법으로 나누어 실험을 했고, 그 결과 누적값의 최소값을 이용하였을 경우 98.33%의 인식률을 얻을 수 있었던 반면에 유전자 알고리듬을 사용하였을 경우 100%의 화자종속 인식률을 얻을 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.695-697
• /
• 2023

최근 몇 년 간 딥러닝 기반 모델의 규모와 복잡성이 증가하면서 강력하고, 높은 정확도가 확보되지만 많은 양의 계산 자원과 메모리가 필요하기 때문에 모바일 장치나 임베디드 시스템과 같은 리소스가 제한된 환경에서의 배포에 제약사항이 생긴다. 복잡한 딥러닝 모델의 배포 및 운영 시 요구되는 고성능 컴퓨터 자원의 문제점을 해결하고자 사전 학습된 대규모 모델로부터 가벼운 모델을 학습시키는 지식증류 기법이 제안되었다. 하지만 현대 딥러닝 기반 모델은 높은 정확도 대비 훈련 데이터에 과적합 되는 과잉 확신(overconfidence) 문제에 대한 대책이 필요하다. 본 논문은 효율적인 경량화를 위한 미리 학습된 모델의 과잉 확신을 방지하고자 초점 손실(focal loss)을 이용한 모델 보정 기법을 언급하며, 다양한 손실 함수 변형에 따라서 지식증류의 성능이 어떻게 변화하는지에 대해 탐구하고자 한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.9C
• /
• pp.905-912
• /
• 2007

This paper proposes an efficient estimation and compensation scheme of IQ imbalance for OFDM-based WLAN systems in the presence of symbol timing error. Since the conventional scheme assumes perfect time synchronization, the criterion of the scheme used to derive the estimation of IQ imbalance is inadequate in the presence of the symbol timing error and the system performance is seriously degraded. New criterion and compensation scheme considering the effect of symbol timing error are proposed. With the proposed scheme, the IQ imbalance can be almost perfectly eliminated in the presence of symbol timing error. The bit error rate performance of the proposed scheme is evaluated by the simulation. In case of 54 Mbps transmission mode in IEEE 802.11a system, the proposed scheme achieves a SNR gain of 4.3dB at $BER=2{\cdot}10^{-3}$. The proposed compensation algorithm of IQ imbalance is implemented using Verilog HDL and verified. The proposed IQ imbalance compensator is composed of 74K logic gates and 6K bits memory from the synthesis result using 0.18um CMOS technology.

• Journal of Internet of Things and Convergence
• /
• v.9 no.1
• /
• pp.9-17
• /
• 2023

Deep learning with large amount of computations is difficult to implement on micro-sized IoT devices or moblie devices. Recently, lightweight deep learning technologies have been introduced to make sure that deep learning can be implemented even on small devices by reducing the amount of computation of the model. Quantization is one of lightweight techniques that can be efficiently used to reduce the memory and size of the model by expressing parameter values with continuous distribution as discrete values of fixed bits. However, the accuracy of the model is reduced due to discrete value representation in quantization. In this paper, we introduce various quantization techniques to correct the accuracy. We selected APoT and EWGS from existing quantization techniques, and comparatively analyzed the results through experimentations The selected techniques were trained and tested with CIFAR-10 or CIFAR-100 datasets in the ResNet model. We found out problems with them through experimental results analysis and presented directions for future research.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.22 no.1
• /
• pp.40-52
• /
• 1997

It has been known that the OFDM-based terrestrial broadcasting systems are significantly more sensitive to nonlinear distortion in high-power amplifiers (HPA's) than the broadcasting systems using single-carrier modulation technique. In this paper, an adaptive precompensator which can compensate the nonlinear effect of the HPA for OEDM systems is proposed. The proposed precompensator consists of a compensation table, realized by memory (RAM), and an adaptive algorithm which can update the values of the compensation table more efficiently using a

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.4
• /
• pp.884-892
• /
• 1996

An adaptive blind equalizer is used to combat the distortions caused by a nonideal channel without resorting to a training sequence, given the received signal and statistical information of the transmitted signal. Incidentally, a multipath channel may result in a fade which produces intersymbol interference in the received signal. Therefore, a new type of algorithm which can compenste the effects of this fade is required in the multipath channel environment. In this paper, a modified form of adaptive blind equalization algorithm using stop-and-go algorithm for multichannel system is proposed. It is demonstrated via computer simulations that the performance of the proposed multichannel stop-and-go algorithm is much better than that of the conventional multichannel algorithms.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.29 no.5
• /
• pp.505-511
• /
• 2023

In the present study, a deep learning model was established to predict the motion response of small fishing vessels. Hydrodynamic performances were evaluated for two small fishing vessels for the dataset of deep learning model. The deep learning model of the Long Short-Term Memory (LSTM) which is one of the recurrent neural network was utilized. The input data of LSTM model consisted of time series of six(6) degrees of freedom motions and wave height and the output label was selected as the time series data of six(6) degrees of freedom motions. The hyperparameter and input window length studies were performed to optimize LSTM model. The time series motion response according to different wave direction was predicted by establised LSTM. The predicted time series motion response showed good overall agreement with the analysis results. As the length of the time series increased, differences between the predicted values and analysis results were increased, which is due to the reduced influence of long-term data in the training process. The overall error of the predicted data indicated that more than 85% of the data showed an error within 10%. The established LSTM model is expected to be utilized in monitoring and alarm systems for small fishing vessels.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.8
• /
• pp.11-21
• /
• 2024

This study explores the use of virtual reality (VR) technology to identify and label key segments of the golf swing. To address the limitations of existing VR devices, we developed a platform to collect kinematic data from various VR devices using the OpenVR SDK (Software Development Kit) and SteamVR, and developed a semi-automated labeling technique to identify and label temporal changes in kinematic behavior through LSTM (Long Short-Term Memory)-based time series data analysis. The experiment consisted of 80 participants, 20 from each of the following age groups: teenage, young-adult, middle-aged, and elderly, collecting data from five swings each to build a total of 400 kinematic datasets. The proposed technique achieved consistently high accuracy (≥0.94) and F1 Score (≥0.95) across all age groups for the seven main phases of the golf swing. This work aims to lay the groundwork for segmenting exercise data and precisely assessing athletic performance on a segment-by-segment basis, thereby providing personalized feedback to individual users during future education and training.