• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.760-762
• /
• 2005

Support vector machine(SVM)은 졸은 일반화 능력을 가진 학습시스템으로, 최근 다양한 패턴 인식 분야에서 적용되고 있다. SVM은 기본적으로 이진 분류기이므로 두 개 이상의 클래스를 분류하기 위해서는 다중 클래스 분류가 가능한 형태로의 설계 방법이 필요하다. 이를 위해 각 클래스별로 독립적인 SVM들을 만들어 결과를 병합하는 방식이 주로 사용되어 왔다. 그러나 이러한 방법은 클래스의 수는 않고 한 클래스 내의 데이터의 수가 많지 않은 경우에는 SVM의 일반화 성능을 저하시키고 노이즈에 민감해지는 문제점을 가지고 있다. 이를 해결하기 위해 본 논문에서는 각 클래스내의 데이터간의 유사도 측정을 위한 통계적 정보를 안정적으로 추출하기 위해 두 데이터의 쌍을 입력으로 받는 새로운 SVM 설계 방법을 제시한다. 제안한 방법을 실제 생체인식 데이터에 적용한 실험에서 기존의 방법보다 우수한 분류 성능을 보임을 확인할 수 있었다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.95-97
• /
• 2019

본 논문에서는 비디오 데이터를 이용한 감독 학습 프레임 워크를 제안한다. 최근 Deep Convolutional Neural Networks의 성공으로 많은 분야에서 사용되고 있다. DCNNs 모델 성능의 중요한 요소 중 하나는 Large-cale Dataset을 구축하는 것으로 Small-scale Dataset으로 모델을 학습한다면 과적합 및 일반화 오류를 해결하기 어렵다. 이러한 문제점을 해결하는 방법으로 이미지 왜곡을 통한 데이터 셋을 증가 또는 Dropout 기법 등을 사용하였지만 원본 데이터가 적은 경우에는 모델이 일반화 능력을 갖기 어렵다. 따라서 본 논문에서는 이러한 문제점을 보완하고자 Web으로부터 얻은 비디오에서 해당 Class와 관련된 프레임들을 추출하여 보다 쉽게 데이터 셋을 확장하고, 모델의 성능을 향상 시키는 방법을 제안한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.553-556
• /
• 2023

데이터 증강은 데이터 불균형 문제를 해결하기 위해 일반화 성능을 향상시킨다. 이는 과적합 문제를 해결하고 정확도를 높이는 데 도움을 준다. 과적합을 해결하기 위해서 본 논문에서는 분할 마스크 라벨링을 자동화하여 효율성을 높이고, RoI를 활용한 분할 Copy-Paste 데이터 증강 기법을 제안한다. 본 논문의 제안 방법을 적용한 결과 YOLOv8 모델에서 기존의 분할, 박스 Copy-Paste 데이터 증강 기법과 비교해서 쓰러진 사람 객체에 대한 정확도가 10.2% 증가함으로써 제안한 방법이 일반화 성능을 높이는 데 효과가 있음을 확인하였다.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.38 no.11
• /
• pp.41-48
• /
• 2001

The exact general bit error rate (TIER) expression of M-ary square quadrature amplitude modulation (QAM) for arbitrary M has not been derived so far. In this paper, a generalized closed-form expression for the BER performance of M-ary square QAM with Gray code bit mapping is derived and analyzed in the presence of additive white Gaussian noise (AWGN) channel. The derivation is based on the consistency of the format in signal constellation o[ Gray coding and it has been derived from the results for M-16, 64, and 256.

• Journal of Korea Multimedia Society
• /
• v.6 no.1
• /
• pp.136-147
• /
• 2003

To achieve an optimal performance for a given problem, we need an integrative process of the parameter optimization via learning and the structure optimization via model selection. In this paper, we propose an efficient optimization method for improving generalization performance by considering the property of each sub-method and by combining them with common theoretical properties. First, weight parameters are optimized by natural gradient teaming with adaptive regularization, which uses a diverse error function. Second, the network structure is optimized by eliminating unnecessary parameters with natural pruning. Through iterating these processes, candidate models are constructed and evaluated based on the Bayesian Information Criterion so that an optimal one is finally selected. Through computational experiments on benchmark problems, we confirm the weight parameter and structure optimization performance of the proposed method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.3C
• /
• pp.237-244
• /
• 2006

I/Q phase unbalance caused by non-ideal circuit components is inevitable physical phenomenons and leads to performance degradation when we implement a practical coherent M-ary phase shift keying(M-PSK) demodulator. In this paper, we present an exact and general expression involving two-dimensional Gaussian Q-functions for the bit error rate(BER) of uniform M-PSK with I/Q phase unbalance over an additive white Gaussian noise(AWGN) channel. First we derive a BER expression for the k-th bit of 8, 16-PSK signal constellations when Gray code bit mapping is employed. Then, from the derived k-th bit BER expression, we present the exact and general average BER expression for M-PSK with I/Q phase unbalance. This result can readily be applied to numerical evaluation for various cases of practical interest in an I/Q unbalanced M-PSK system, because the one- and two-dimensional Gaussian Q-functions can be easily and directly computed using commonly available mathematical software tools.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.5
• /
• pp.1044-1051
• /
• 2003

This paper surveys the EBP(Error Back Propagation) learning, the Cross Entropy function and the LBL(Layer By Layer) learning, which are used for learning the MLP(Multi Layer Perceptrons). We compare the merits and demerits of each learning method in the handwritten digit recognition. Although the speed of EBP learning is slower than other learning methods in the initial learning process, its generalization capability is better. Also, the speed of Cross Entropy function that makes up for the weak points of EBP learning is faster than that of EBP learning. But its generalization capability is worse because the error signal of the output layer trains the target vector linearly. The speed of LBL learning is the fastest speed among the other learning methods in the initial learning process. However, it can't train for more after a certain time, it has the lowest generalization capability. Therefore, this paper proposes the standard of selecting the learning method when we apply the MLP.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.5
• /
• pp.389-394
• /
• 2016

MIMO (Multiple-Input-Multiple-Output) in underwater acoustic communication has distortion of received signal because of ISI (Inter-Symbol Interference) and crosstalk among transmitters. Time-reversal mirror was used for compensating of signal distortion, but it has a limit in eliminating crosstalk effectively. This paper proposes a time-reversal mirror based on GSC (Generalized Sidelobe Canceller) for removing crosstalk. The FAF05 (The Focused Acoustic Forecasting 05) experimental data has been used to verify the suggested method by comparison with the conventional time-reversal for communication performance, and it is demonstrated that the suggested method produces better communication performance results than conventional time-reversal.

• Journal of KIISE:Software and Applications
• /
• v.28 no.9
• /
• pp.603-611
• /
• 2001

This paper proposes a novel fast layer-by-layer algorithm that has better generalization capability. In the proposed algorithm, the weights of the hidden layer are updated by the target vector of the hidden layer obtained by least squares method. The proposed algorithm improves the learning speed that can occur due to the small magnitude of the gradient vector in the hidden layer. This algorithm was tested in a handwritten digits recognition problem. The learning speed of the proposed algorithm was faster than those of error back propagation algorithm and modified error function algorithm, and similar to those of Ooyen's method and layer-by-layer algorithm. Moreover, the simulation results showed that the proposed algorithm had the best generalization capability among them regardless of the number of hidden nodes. The proposed algorithm has the advantages of the learning speed of layer-by-layer algorithm and the generalization capability of error back propagation algorithm and modified error function algorithm.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.5
• /
• pp.509-514
• /
• 2015

In this paper, we introduce a new pattern classifier which is based on the learning algorithm of Extreme Learning Machine the sort of artificial neural networks and fuzzy set theory which is well known as being robust to noise. The learning algorithm used in Extreme Learning Machine is faster than the conventional artificial neural networks. The key advantage of Extreme Learning Machine is the generalization ability for regression problem and classification problem. In order to evaluate the classification ability of the proposed pattern classifier, we make experiments with several machine learning data sets.