• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권1호
• /
• pp.368-391
• /
• 2018

Face recognition (FR) with a single sample per person (SSPP) is common in real-world face recognition applications. In this scenario, it is hard to predict intra-class variations of query samples by gallery samples due to the lack of sufficient training samples. Inspired by the fact that similar faces have similar intra-class variations, we propose a virtual sample generating algorithm called k nearest neighbors based virtual sample generating (kNNVSG) to enrich intra-class variation information for training samples. Furthermore, in order to use the intra-class variation information of the virtual samples generated by kNNVSG algorithm, we propose image set based multimanifold discriminant learning (ISMMDL) algorithm. For ISMMDL algorithm, it learns a projection matrix for each manifold modeled by the local patches of the images of each class, which aims to minimize the margins of intra-manifold and maximize the margins of inter-manifold simultaneously in low-dimensional feature space. Finally, by comprehensively using kNNVSG and ISMMDL algorithms, we propose k nearest neighbor virtual image set based multimanifold discriminant learning (kNNMMDL) approach for single sample face recognition (SSFR) tasks. Experimental results on AR, Multi-PIE and LFW face datasets demonstrate that our approach has promising abilities for SSFR with expression, illumination and disguise variations.

• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.843-848
• /
• 1999

Support vector machine(SVM) is a new and very promising classification and regression technique developed by Bapnik and his group at AT&T Bell laboratories. However it has failed to establish itself as common machine learning tool. This is partly due to the fact that SVM is not easy to implement and its standard implementation requires the optimization package for quadratic programming. In this paper we present simple iterative Kernl Adatron algorithm for nonparametric regression which is easy to implement and guaranteed to converge to the optimal solution and compare it with neural networks and projection pursuit regression.

• 정보저장시스템학회논문집
• /
• 제3권4호
• /
• pp.167-172
• /
• 2007

The recently-developed off-line learning control approaches for the rejection of periodic disturbances utilize the specific property that the learning system tends to oscillate in steady state. Unfortunately, the prior works have not clarified how closely the learning system should approach the steady state to achieve the rejection of periodic disturbances to satisfactory level. In this paper, we address this issue extensively for the class of linear systems. We also attempt to remove the effect of other aperiodic disturbances on the rejection of the periodic disturbances effectively. In fact, the proposed learning control algorithm can provide very fast convergence performance in the presence of aperiodic disturbance. The effectiveness and practicality of our work is demonstrated through mathematical performance analysis as well as various simulation results.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2007년도 심포지엄 논문집 정보 및 제어부문
• /
• pp.107-109
• /
• 2007

The recently-developed off-line learning control approaches for the rejection of periodic disturbances utilize the specific property that the learning system tends to oscillate in steady state. Unfortunately, the prior works have not clarified how closely the learning system should approach the steady state to achieve the rejection of periodic disturbances to satisfactory level. In this paper, we address this issue extensively for the class of linear systems. We also attempt to remove the effect of other aperiodic disturbances on the rejection of the periodic disturbances effectively. In fact, the proposed learning control algorithm can provide very fast convergence performance in the presence of aperiodic disturbance. The effectiveness and practicality of our work is demonstrated through mathematical, performance analysis as well as various simulation results.

• 한국정보과학회논문지:컴퓨팅의 실제 및 레터
• /
• 제8권3호
• /
• pp.344-355
• /
• 2002

MLP(multilayer perceptron)는 다른 패턴인식 방법에 비해 여러 가지 훌륭한 특성을 가지고 있어 패턴인식에서 폭넓게 사용되고 있다. 그러나 MLP의 학습에 일반적으로 사용되는 EBP(error backpropagation) 알고리즘은 학습시간이 비교적 오래 걸린다는 단점이 있다. 패턴인식에 사용되는 학습 데이타는 풍부한 중복특성을 내포하고 있으므로 패턴마다 MLP의 내부변수를 갱신하는 온라인 계열의 학습방식이 속도의 향상에 상당한 효과가 있다. 일반적인 온라인 EBP 알고리즘에서는 내부변수 갱신시 고정된 학습률을 적용한다. 고정 학습률을 적절히 선택함으로써 패턴인식 웅용에서 상당한 속도개선을 얻을 수 있지만, 학습률이 고정되고 학습이 진행됨에 따라 학습패턴 영역이 달라지는 학습과정의 각 단계에 효과적으로 대웅하지 못하는 문제가 있다. 이 문제에 대해 본 논문에서는 학습과정을 세 단계로 정의하고, 각 단계별로 필요한 패턴만을 학습에 반영하는 패턴별 가변학습속도 및 학습생략(ILVRS) 방법을 제안한다. ILVRS의 기본개념은 다음과 같다. 학습단계마다 학습에 필요한 패턴의 부분이 달라지므로 이를 구별 하여 학습에 적용할 수 있도록 (1)패턴마다 발생하는 오류치를 적절한 범위 이내로 제한하여 가변 학습률로 사용하고, (2)학습이 진행됨에 따라 불필요한 부분의 패턴을 학습에서 생략한다. 제안한 ILVRS의 성능을 입증하기 위해 본 논문에서는 패턴인식 응용의 한 갈래인 화자증명을 실험하고 그 결과를 제시한다.

• 대한임베디드공학회논문지
• /
• 제14권5호
• /
• pp.277-286
• /
• 2019

In this paper, we propose an efficient hyperplane generation technique to classify human activity from combination of events and sequence information obtained from multiple-event sensors. By generating hyperplane efficiently, our machine learning algorithm classify with less memory and run time than the LSVM (Linear Support Vector Machine) for embedded system. Because the fact that light weight and high speed algorithm is one of the most critical issue in the IoT, the study can be applied to smart home to predict human activity and provide related services. Our approach is based on reducing numbers of hyperplanes and utilizing robust string comparing algorithm. The proposed method results in reduction of memory consumption compared to the conventional ML (Machine Learning) algorithms; 252 times to LSVM and 34,033 times to LSTM (Long Short-Term Memory), although accuracy is decreased slightly. Thus our method showed outstanding performance on accuracy per hyperplane; 240 times to LSVM and 30,520 times to LSTM. The binarized image is then divided into groups, where each groups are converted to binary number, in order to reduce the number of comparison done in runtime process. The binary numbers are then converted to string. The test data is evaluated by converting to string and measuring similarity between hyperplanes using Levenshtein algorithm, which is a robust dynamic string comparing algorithm. This technique reduces runtime and enables the proposed algorithm to become 27% faster than LSVM, and 90% faster than LSTM.

• 전기전자학회논문지
• /
• 제12권1호
• /
• pp.65-74
• /
• 2008

패턴 분류에 많이 사용되는 기법 중의 하나인 메모리 기반 추론 알고리즘은 단순히 메모리에 저장하고 분류 시에 저장된 패턴과 테스트 패턴간의 거리를 계산하여 가장 가까운 학습패턴의 클래스로 분류하는 기법이기 때문에 패턴의 개수가 늘어나면 메모리가 증가하고 또한 추가로 패턴이 발생할 경우 처음부터 다시 수행해야하는 문제점을 가지고 있다. 이러한 문제점을 해결하기 위하여 이미 학습한 대표패턴을 기억하고 새로 들어오는 패턴에 대해서만 학습하는 점진적 학습 방법을 제안한다. 즉 추가로 학습패턴이 발생할 경우 매번 전체 학습 패턴을 다시 학습하는 것이 아니라, 새로 추가된 데이터만을 학습하여 대표패턴을 추출하여 메모리사용을 줄이는 iMPA(incremental Multi Partition Averaging)기법을 제안하였다. 본 논문에서 제안한 기법은 대표적인 메모리기반 추론 기법인 k-NN 기법과 비교하여 현저하게 줄어든 대표패턴으로 유사한 분류 성능을 보여주며, 점진적 특성을 지닌 NGE 이론을 구현한 EACH 시스템과 점진적인 실험에서도 탁월한 분류 성능을 보여준다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권12호
• /
• pp.4308-4325
• /
• 2021

The importance and necessity of artificial intelligence, particularly machine learning, has recently been emphasized. In fact, artificial intelligence, such as intelligent surveillance cameras and other security systems, is used to solve various problems or provide convenience, providing solutions to problems that humans traditionally had to manually deal with one at a time. Among them, information security is one of the domains where the use of artificial intelligence is especially needed because the frequency of occurrence and processing capacity of dangerous codes exceeds the capabilities of humans. Therefore, this study intends to examine the definition of artificial intelligence and machine learning, its execution method, process, learning algorithm, and cases of utilization in various domains, particularly the cases and contents of artificial intelligence technology used in the field of information security. Based on this, this study proposes a method to apply machine learning technology to the method of classifying and detecting malware that has rapidly increased in recent years. The proposed methodology converts software programs containing malicious codes into images and creates training data suitable for machine learning by preparing data and augmenting the dataset. The model trained using the images created in this manner is expected to be effective in classifying and detecting malware.

• 반도체디스플레이기술학회지
• /
• 제17권1호
• /
• pp.71-75
• /
• 2018

Image enhancement is a necessary end essential step after taking a picture with a digital camera. Many different photo software packages attempt to automate this process with various auto enhancement techniques. This paper provides and implements a system that can learn a user's preferences and apply the preferences into the process of image enhancement. Five major components are applied to the implemented system, which are computing a distance metric, finding a training set, finding an optimal parameter set, training and finally enhancing the input image. To estimate the validity of the method, we carried out user studies, and the fact that the implemented system was preferred over the method without learning user preferences.

• 정보교육학회논문지
• /
• 제11권1호
• /
• pp.29-38
• /
• 2007

최근 컴퓨터 교육에서 프로그래밍에 대한 관심이 높아지고 있다. 학습에서 논리력과 창의력의 요구가 늘어나고 있는 추세에 힘입어 컴퓨터 교육 분야에서도 프로그래밍 학습을 통한 논리력과 창의력을 함양하려는 다양한 시도가 이루어지고 있다. 프로그램은 알고리즘을 컴퓨터 언어로 구현한 것이므로 문제해결의 핵심은 논리적 해결 방법의 고안 즉, 알고리즘의 고안에 있다. 알고리즘의 중요성을 인식한다면 이제 알고리즘 교육에 대한 방향 정립을 위한 체계적 연구가 필요한 시점에 와 있다고 볼 수 있다. 지금까지 주로 정렬, 탐색과 같은 특정 문제를 중심으로 한 문제 종속적이고 개별적인 알고리즘 학습을 강조해 왔다. 본 논문에서는 문제 독립적, 집단적 성질을 고려하는 문제 해결의 개념적 유형이라는 관점에서 개념적 알고리즘(conceptual algorithm)을 정의한다. 개념적 알고리즘의 유형적 특성을 바탕으로 개별적, 문제 종속적인 기존의 알고리즘 교육과 구분되는 체계적, 문제 독립적, 알고리즘적인 문제 해결력을 지향하는 알고리즘 교육의 새로운 방향을 제안한다.