• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권11호
• /
• pp.2924-2944
• /
• 2023

Alzheimer's disease (AD) is a neurological condition that is recognized as one of the primary causes of memory loss. AD currently has no cure. Therefore, the need to develop an efficient model with high precision for timely detection of the disease is very essential. When AD is detected early, treatment would be most likely successful. The most often utilized indicators for AD identification are the Mini-mental state examination (MMSE), and the clinical dementia. However, the use of these indicators as ground truth marking could be imprecise for AD detection. Researchers have proposed several computer-aided frameworks and lately, the supervised model is mostly used. In this study, we propose a novel 3D Convolutional Neural Network Multilayer Perceptron (3D CNN-MLP) based model for AD classification. The model uses Attention Mechanism to automatically extract relevant features from Magnetic Resonance Images (MRI) to generate probability maps which serves as input for the MLP classifier. Three MRI scan categories were considered, thus AD dementia patients, Mild Cognitive Impairment patients (MCI), and Normal Control (NC) or healthy patients. The performance of the model is assessed by comparing basic CNN, VGG16, DenseNet models, and other state of the art works. The models were adjusted to fit the 3D images before the comparison was done. Our model exhibited excellent classification performance, with an accuracy of 91.27% for AD and NC, 80.85% for MCI and NC, and 87.34% for AD and MCI.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제36권8호
• /
• pp.633-641
• /
• 2009

본 논문에서는 아날로그 데이터 처리가 가능하고, 온라인 학습, 학습 중 새로운 클래스 추가등의 특징을 가진 패턴 인식기를 제안하였다. 제안한 패턴 인식기는 계층적 구조를 가지고 있으며, 각 레벨별로 서로 다른 메트릭을 적용하여 분류 성능을 향상 시켰다. 제안한 패턴 인식기는 신경망 기반의 패턴 인식 알고리즘인 Gaussian ARTMAP 모델을 기반으로 하고 있다. Gaussian ARTMAP 모델을 계층적으로 구성하고, 계층마다 서로 다른 특징을 학습하도록 하기 위하여 Principal Component Emphasis (P.C.E) 방법을 제안하였으며, 이를 이용하여 새로운 메트릭을 생성하는 방법을 제안하였다. P.C.E는 학습된 입력 데이터들의 분산을 이용하여 클래스 내의 공통 속성을 나타내는 분산이 작은 차원을 제거하고 패턴 간의 서로 다른 속성을 나타내는 분산이 큰 차원만 유지하는 방법이다. 제안한 알고리즘의 학습 과정에서 교사 신호와 다르게 분류된 패턴이 발생하면 잘못 분류 된 클래스와 입력된 패턴을 분리하기 위하여 P.C.E를 수행하고 하위 노드에서 학습하게 된다. 실험 결과 제안한 모델은 기존에 제안된 패턴 인식 모델들 보다 높은 분류 성능을 가지고 있음을 확인하였다.

• 전기학회논문지
• /
• 제66권4호
• /
• pp.682-691
• /
• 2017

In this paper, we propose the fusion design methodology of both pedestrian detection and object tracking system realized with the aid of HOG-PCA based RBFNN pattern classifier. The proposed system includes detection and tracking parts. In the detection part, HOG features are extracted from input images for pedestrian detection. Dimension reduction is also dealt with in order to improve detection performance as well as processing speed by using PCA which is known as a typical dimension reduction method. The reduced features can be used as the input of the FCM-based RBFNNs pattern classifier to carry out the pedestrian detection. FCM-based RBFNNs pattern classifier consists of condition, conclusion, and inference parts. FCM clustering algorithm is used as the activation function of hidden layer. In the conclusion part of network, polynomial functions such as constant, linear, quadratic and modified quadratic are regarded as connection weights and their coefficients of polynomial function are estimated by LSE-based learning. In the tracking part, object tracking algorithms such as mean shift(MS) and cam shift(CS) leads to trace one of the pedestrian candidates nominated in the detection part. Finally, INRIA person database is used in order to evaluate the performance of the pedestrian detection of the proposed system while MIT pedestrian video as well as indoor and outdoor videos obtained from IC&CI laboratory in Suwon University are exploited to evaluate the performance of tracking.

• 지능정보연구
• /
• 제28권3호
• /
• pp.119-137
• /
• 2022

최근 추천 시스템 연구에서는 사용자와 아이템 간 상호 작용을 보다 잘 표현하고자 다양한 딥 러닝 모델을 적용하고 있다. ONCF(Outer product-based Neural Collaborative Filtering)는 사용자와 아이템의 행렬을 외적하고 합성곱 신경망을 거치는 구조로 2차원 상호작용 맵을 제작해 사용자와 아이템 간의 상호 작용을 더욱 잘 포착하고자 한 대표적인 딥러닝 기반 추천시스템이다. 하지만 합성곱 신경망을 이용하는 ONCF는 학습 데이터에 나타나지 않은 분포를 갖는 데이터의 경우 예측성능이 떨어지는 귀납적 편향을 가지는 한계가 있다. 본 연구에서는 먼저 NCF구조에 Transformer에 기반한 ViT(Vision Transformer)를 도입한 방법론을 제안한다. ViT는 NLP분야에서 주로 사용되던 트랜스포머를 이미지 분류에 적용하여 좋은 성과를 거둔 방법으로 귀납적 편향이 합성곱 신경망보다 약해 처음 보는 분포에도 robust한 특징이 있다. 다음으로, ONCF는 사용자와 아이템에 대한 단일 잠재 벡터를 사용하였지만 본 연구에서는 모델이 더욱 다채로운 표현을 학습하고 앙상블 효과도 얻기 위해 잠재 벡터를 여러 개 사용하여 채널을 구성한다. 마지막으로 ONCF와 달리 부가 정보(side information)를 추천에 반영할 수 있는 아키텍처를 제시한다. 단순한 입력 결합 방식을 활용하여 신경망에 부가 정보를 반영하는 기존 연구와 달리 본 연구에서는 독립적인 보조 분류기(auxiliary classifier)를 도입하여 추천 시스템에 부가정보를 보다 효율적으로 반영할 수 있도록 하였다. 결론적으로 본 논문에서는 ViT 의 적용, 임베딩 벡터의 채널화, 부가정보 분류기의 도입을 적용한 새로운 딥러닝 모델을 제안하였으며 실험 결과 ONCF보다 높은 성능을 보였다.

• 한국음향학회지
• /
• 제33권4호
• /
• pp.268-272
• /
• 2014

본 논문에서는 MSVQ(Multi-Section Vector Quantization)와 시간지연 회귀 신경회로망(TDRNN)을 이용한 하이브리드 구조의 음성인식 방법을 제안한다. MSVQ는 음성의 길이를 일정한 구간 수로 정규화한 코드북을 생성하고, 시간지연 회귀 신경회로망은 이 코드북을 이용하여 음성을 인식한다. 시간지연 회귀 신경회로망은 음성의 시계열 문맥정보를 잘 학습할 수 있는 구조로 구성되었다. 음성특징으로 인지선형예측(PLP) 계수가 사용되었다. 음성인식 실험을 수행한 결과 MSVQ/TDRNN 음성인식기는 97.9 %의 화자독립 음성 인식률을 보였다.

• 한국산학기술학회논문지
• /
• 제3권4호
• /
• pp.245-250
• /
• 2002

본 연구에서는 영문 단어로부터 폰트를 분류하기 위해 연역적이고 국부적인 폰트 분류 방법을 제안한다. 이는 문자 인식 전에 한 단어의 폰트를 분류하는 것을 말한다. 폰트 분류를 위해 활자 특성인 Ascender, Descender와 Serif가 사용된다. 입력 단어로부터 Ascender, Descender 와 Serif가 추출되어 경사도 특징 벡터가 추출되고, 그 특징 벡터는 인공 신경망에 의해 입력 단어에 대한 폰트 스타일, 폰트 그룹, 폰트 이름이 분류된다. 제안된 연역적이고 국부적인 폰트 분류 방법은 폰트 정보가 문자 분할기와 문자 인식기에 사용될 수 있게 한다. 나아가, 특정 폰트에 따른 Mono-Font 문자 분할기와 Mono-Font 문자 인식기로 구성되는 OCR 시스템을 구성할 수 있는 것을 가능하게 한다.

• 제어로봇시스템학회논문지
• /
• 제14권8호
• /
• pp.818-823
• /
• 2008

In this paper, we propose a feature extraction method which extracts directional features of handwritten numerals by using the projection runlength. Our directional featrures are obtained from four directional images, each of which contains horizontal, vertical, right-diagonal and left-diagonal lines in entire numeral shape respectively. A conventional method which extracts directional features by using Kirsch masks generates edge-shaped double line directional images for four directions, whereas our method uses the projections and their runlengths for four directions to produces single line directional images for four directions. To obtain the directional projections for four directions from a numeral image, some preprocessing steps such as thinning and dilation are required, but the shapes of resultant directional lines are more similar to the numeral lines of input numerals. Four [$4{\times}4$] directional features of a numeral are obtained from four directional line images through a zoning method. By using a hybrid feature which is made by combining our feature with the conventional features of a mesh features, a kirsch directional feature and a concavity feature, higher recognition rates of the handwrittern numerals can be obtained. For recognition test with given features, we use a multi-layer perceptron neural network classifier which is trained with the back propagation algorithm. Through the experiments with the handwritten numeral database of Concordia University, we have achieved a recognition rate of 97.85%.

• Journal of Ship and Ocean Technology
• /
• 제8권2호
• /
• pp.41-60
• /
• 2004

Ship design process requires lots of complicated analyses for determining a large number of design variables. Due to its complexity, the process is divided into several tractable designs or analysis problems. The interdependent relationship requires repetitive works. This paper employs collaborative optimization (CO), one of the multidisciplinary design optimization (MDO) techniques, for treating such complex relationship. CO guarantees disciplinary autonomy while maintaining interdisciplinary compatibility due to its bi-level optimization structure. However, the considerably increased computational time and the slow convergence have been reported as its drawbacks. This paper proposes the use of an approximation model in place of the disciplinary optimization in the system-level optimization. Neural network classification is employed as a classifier to determine whether a design point is feasible or not. Kriging is also combined with the classification to make up for the weakness that the classification cannot estimate the degree of infeasibility. For the purpose of enhancing the accuracy of a predicted optimum and reducing the required number of disciplinary optimizations, an approximation management framework is also employed in the system-level optimization.

• 대한방사선기술학회지:방사선기술과학
• /
• 제43권3호
• /
• pp.195-208
• /
• 2020

Artificial intelligence(AI) is a field of computer science that is defined as allowing computers to imitate human intellectual behavior, even though AI's performance is to imitate humans. It is grafted across software-based fields with the advantages of high accuracy and speed of processing that surpasses humans. Indeed, the AI based technology has become a key technology in the medical field that will lead the development of medical image analysis. Therefore, this article introduces and discusses the concept of deep learning-based medical imaging analysis using the principle of algorithms for convolutional neural network(CNN) and back propagation. The research cases application of the AI based medical imaging analysis is used to classify the various disease(such as chest disease, coronary artery disease, and cerebrovascular disease), and the performance estimation comparing between AI based medical imaging classifier and human experts.

• 전기학회논문지
• /
• 제57권12호
• /
• pp.2313-2321
• /
• 2008

In order to develop reliable on-site partial discharge(PD) pattern recognition algorithm, the fuzzy neural network based on fuzzy set(FNN) and the polynomial network pattern classifier based on fuzzy Inference(PNC) were investigated and designed. Using PD data measured from laboratory defect models, these algorithms were learned and tested. Considering on-site situation where it is not easy to obtain voltage phases in PRPDA(Phase Resolved Partial Discharge Analysis), the measured PD data were artificially changed with shifted voltage phases for the test of the proposed algorithms. As input vectors of the algorithms, PRPD data themselves were adopted instead of using statistical parameters such as skewness and kurtotis, to improve uncertainty of statistical parameters, even though the number of input vectors were considerably increased. Also, results of the proposed neuro-fuzzy algorithms were compared with that of conventional BP-NN(Back Propagation Neural Networks) algorithm using the same data. The FNN and PNC algorithms proposed in this study were appeared to have better performance than BP-NN algorithm.