• Journal of Information Processing Systems
• /
• 제15권4호
• /
• pp.957-966
• /
• 2019

This paper studies a novel approach to natural gait cycles based gait recognition via kernel Fisher discriminant analysis (KFDA), which can effectively calculate the features from gait sequences and accelerate the recognition process. The proposed approach firstly extracts the gait silhouettes through moving object detection and segmentation from each gait videos. Secondly, gait energy images (GEIs) are calculated for each gait videos, and used as gait features. Thirdly, KFDA method is used to refine the extracted gait features, and low-dimensional feature vectors for each gait videos can be got. The last is the nearest neighbor classifier is applied to classify. The proposed method is evaluated on the CASIA and USF gait databases, and the results show that our proposed algorithm can get better recognition effect than other existing algorithms.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2005년도 Proceedings of ISRS 2005
• /
• pp.209-212
• /
• 2005

Hyperspectral data consists of more than 200 spectral bands that are highly correlated. In order to utilize hyperspectral data for classification, dimensional reduction or feature extraction is desired. By applying feature extraction, computational complexity of classification can be reduced and classification accuracy may be improved. In this paper, a genetic algorithm based feature extraction technique is proposed. Measure from discriminant analysis is utilized as optimization criterion. A subset of spectral bands is selected by genetic algorithm. Dimension of feature space is further reduced by linear transformation. Feasibility of the proposed technique is evaluated with AVIRIS data.

• Journal of Information Technology Applications and Management
• /
• 제13권1호
• /
• pp.77-94
• /
• 2006

Case-based Reasoning (CBR) is widely used in business and industry prediction. It is suitable to solve complex and unstructured business problems. Recently, the prediction accuracy of CBR has been enhanced by not only various machine learning algorithms such as genetic algorithms, relative weighting of Artificial Neural Network (ANN) input variable but also data mining technique such as feature selection, feature weighting, feature transformation, and instance selection As a result, CBR is even more widely used today in business area. In this study, we investigated the usefulness of the CBR method in forecasting success in implementing ERP systems. We used a CBR method based on the feature weighting technique to compare the performance of three different models : MDA (Multiple Discriminant Analysis), GECBR (GEneral CBR), FWCBR (CBR with Feature Weighting supported by Analytic Hierarchy Process). The study suggests that the FWCBR approach is a promising method for forecasting of successful ERP implementation in Small and Medium sized Enterprises.

• 디지털융복합연구
• /
• 제11권12호
• /
• pp.295-302
• /
• 2013

사람 검출은 정지된 영상 혹은 동영상으로부터 사람의 움직임이나 자세를 추정하고, 사람이 찾아질 경우 영상 내 사람의 좌표, 동작 인식, 보안관련 인증 등을 알아내는 기술로 정의된다. 이러한 사람 검출은 다른 객체의 검출이나 사람과 컴퓨터와의 상호작용, 동작 인식 등의 기초 기술로서 해당 시스템의 성능에 영향을 미치는 매우 중요한 변수 중에 하나이다. 그러나 영상 내의 사람은 움직임, 자세, 크기, 빛의 방향 및 밝기, 다른 객체와의 중복 등의 환경적 변화로 인해 사람 모양이 다양해지므로 정확하고 빠른 검출이 어렵다. 따라서 본 논문에서는 피셔의 선형 판별 분석을 이용하여 몇 가지 환경적 조건을 극복한 정확하고 빠른 사람 검출 방법을 제안한다. 제안된 방법은 사람 움직임 및 자세와 배경에 무관하게 빠른 시간 안에 사람을 검출하는 것이 가능하다. 이를 위해 계층적인 방법으로 사람 검출을 수행하며, 휴리스틱한 방법, 피셔의 판별 분석을 이용하여 사람 검출을 수행하고, 검색 영역의 축소와 선형 결정의 계산 시간의 단축으로 검출 응답 시간을 빠르게 하였다. 추출된 사람 영상에서 사람의 자세를 추정하고 사람의 영역을 검출함으로써 사람 정보의 사용에 있어 보다 많은 정보를 추출할 수 있도록 하였다.

• 대한전자공학회논문지SP
• /
• 제43권4호
• /
• pp.87-95
• /
• 2006

본 논문에서는 얼굴 영상의 에너지 분포 특성을 이용한 새로운 특정추출 방법을 제안한다. 제안된 방법은 얼굴 영상의 에너지 확률과 에너지 랩을 이용해서 데이터 차원이 축소된 유효정보의 추출 및 유효정보의 LDA 해석에 기반을 둔다. 일반적으로, 얼굴 영상은 고유한 에너지 분포 특성을 가지고 있다. 그러나 기존의 많은 DCT 기반 방법들은 이러한 얼굴 영상의 특성을 효과적으로 이용하지 못하는 단점이 있다. 제안된 방법은 이러한 기존 방법의 단점을 개선하기 위해 다음의 3단계 방법을 사용한다. 먼저, DCT 도메인에서 얼굴의 에너지 확률 개념을 정의하고, 이러한 에너지 확률로부터 얼굴의 에너지 맵을 생성한다. 마지막으로, 에너지 확률 지도에 위치한 주파수 계수들에 대한 LDA 적용 및 해석을 통하여 특정 벡터 추출 및 인식을 수행한다. 제안된 방법은 ETRI 데이터베이스에서 96.8%, ORL 데이터베이스에서 100%의 인식률을 보인다. 실험을 통하여 인식 성능의 개선뿐만 아니라, 특정 벡터의 차원 축소에도 효과가 있음을 알 수 있다.

• 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.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권11호
• /
• pp.4502-4518
• /
• 2015

Fisher linear discriminant analysis (LDA) is one of the most popular projection techniques for feature extraction and has been widely applied in face recognition. However, it cannot be used when encountering the single sample per person problem (SSPP) because the intra-class variations cannot be evaluated. In this paper, we propose a novel method called local similarity based linear discriminant analysis (LS_LDA) to solve this problem. Motivated by the "divide-conquer" strategy, we first divide the face into local blocks, and classify each local block, and then integrate all the classification results to make final decision. To make LDA feasible for SSPP problem, we further divide each block into overlapped patches and assume that these patches are from the same class. To improve the robustness of LS_LDA to outliers, we further propose local similarity based median discriminant analysis (LS_MDA), which uses class median vector to estimate the class population mean in LDA modeling. Experimental results on three popular databases show that our methods not only generalize well SSPP problem but also have strong robustness to expression, illumination, occlusion and time variation.

• 정보보호학회논문지
• /
• 제24권6호
• /
• pp.1055-1063
• /
• 2014

전력소모량을 이용한 부채널 분석의 성능 향상을 위해 다양한 분석기법이 제안되고 있다. 이들 중, 사전처리 단계에서 적용 가능한 파형압축은 전력분석을 위한 소요시간을 단축하고 수집신호의 잡음성분을 줄이기 위해 널리 사용되는 방법이다. 본 논문에서는 영상처리 등에 많이 사용되고 있는 선형판별분석(Linear Discriminant Analysis)을 이용한 전력분석기법을 제안한다. 또한, 실험을 통해 기존의 파형압축방법 중 가장 성능이 좋은 것으로 알려진 주성분분석(Principal Component Analysis)을 이용한 방법과의 성능 비교를 통해 제안기법의 우수성을 증명한다.

• ETRI Journal
• /
• 제36권1호
• /
• pp.99-105
• /
• 2014

This paper presents a new locomotion mode recognition method based on a transformed correlation feature analysis using an electromyography (EMG) pattern. Each movement is recognized using six weighted subcorrelation filters, which are applied to the correlation feature analysis through the use of six time-domain features. The proposed method has a high recognition rate because it reflects the importance of the different features according to the movements and thereby enables one to recognize real-time EMG patterns, owing to the rapid execution of the correlation feature analysis. The experiment results show that the discriminating power of the proposed method is 85.89% (${\pm}2.5$) when walking on a level surface, 96.47% (${\pm}0.9$) when going up stairs, and 96.37% (${\pm}1.3$) when going down stairs for given normal movement data. This makes its accuracy and stability better than that found for the principal component analysis and linear discriminant analysis methods.

• 한국생산제조학회지
• /
• 제9권1호
• /
• pp.119-127
• /
• 2000

In this study the classified researches the artificial and natural flaws in welding parts are performed using the pattern recognition technology. For this purpose the signal pattern recognition package including the user defined function was developed and the total procedure including the digital signal processing feature extraction feature selection and classifi-er selection is teated by bulk,. Specially it is composed with and discussed using the statistical classifier such as the linear discriminant function the empirical Bayesian classifier. Also the pattern recognition technology is applied to classifica-tion problem of natural flaw(i.e multiple classification problem-crack lack of penetration lack of fusion porosity and slag inclusion the planar and volumetric flaw classification problem), According to this result it is possible to acquire the recognition rate of 83% above even through it is different a little according to domain extracting the feature and the classifier.