• 한국멀티미디어학회논문지
• /
• 제18권7호
• /
• pp.834-842
• /
• 2015

We propose a method to construct composite feature vector based on discriminant analysis for face recognition. For this, we first extract the holistic- and local-features from whole face images and local images, which consist of the discriminant pixels, by using a discriminant feature extraction method. In order to utilize both advantages of holistic- and local-features, we evaluate the amount of the discriminative information in each feature and then construct a composite feature vector with only the features that contain a large amount of discriminative information. The experimental results for the FERET, CMU-PIE and Yale B databases show that the proposed composite feature vector has improvement of face recognition performance.

• 지능정보연구
• /
• 제2권1호
• /
• pp.45-58
• /
• 1996

This paper describes an a, pp.oach to recognizing composite features of prismatic parts. AAG (Attribute Adjacency Graph) is adopted as the basis of describing basic feature, but it is extended to enhance the expressive power of AAG by adding face type, angles between faces and normal vectors. Our a, pp.oach is called Extended AAG (EAAG). To simplify the recognition procedure, feature classification tree is built using the graph types of EEA and the number of EAD's. Algorithms to find open faces and dimensions of features are exemplified and used in decomposing composite feature. The processing sequence of recognized features is automatically determined during the decomposition process of composite features.

• 한국CDE학회논문집
• /
• 제3권4호
• /
• pp.274-282
• /
• 1998

A feature is a local shape of a product directly related to the manufacturing process. The feature plays a role of the bridge connecting CAD and CAM. In the process planning for he CAM, information on manufacturing is required. To get the a manufacturing information from CAD dat, we need to recognize features. Once features are recognized, they are used as an input for the process planning. In this paper, we thoroughly investigate the composite features, which are generated by interacting simple features. The simple features in the composite feature usually have precedence relation in terms of process sequence. Based on the reason for the precedence relation, we classify the composite features for the process planning. In addition to the precedence relation, approach direction is used as an input for the process planning. In the process planning, the number of set-up orientations are minimized whole process sequence for the features are generated. We propose a process planning algorithm based on the topological sort and breadth-first search of graphs. The algorithn is verified using sample products.

• 한국정보통신학회논문지
• /
• 제14권12호
• /
• pp.2761-2768
• /
• 2010

본 논문에서는 숫자의 전경특징과 배경특징을 이용하고 SVM 분류기를 사용하여 오프라인 필기체 숫자인식에서 인식률을 향상시키는 방안을 제시한다. 숫자의 전경특징은 숫자의 에지선을 추출한 Kirsch 방향특징과 숫자선 자체를 추출한 projection 방향특징으로 구성되며, 숫자의 배경특징은 숫자의 볼록외피로 부터 추출되는 오목특징이다. 여기서 오목특징은 방향특징에 대해 보완적인 특징으로 작용하여 분류 성능 향상에 기여한다. 인식기로는 RBF 커널을 이용한 SVM 분류기를 사용하고, CENPAMI 숫자특징 데이터베이스를 사용하여 제시된 방법의 성능을 검사하였다. 실험 결과 각기 다른 분류 성능을 갖는 이들 3종의 특징들이 상호 보완적으로 작용하여 인식률 향상에 기여함을 확인할 수 있었으며, 제시된 복합특징에 의해 98.90%의 인식률을 달성하였다.

• 전자공학회논문지
• /
• 제52권10호
• /
• pp.129-136
• /
• 2015

본 논문에서는 조명 변이에 강인하게 동작 할 수 있는 하이브리드 얼굴 인식 방법을 제안한다. 이를 위해, 서로 다른 특성을 가진 조명 불변 특징 추출 방법으로부터 판별력 있는 특징들을 추출한다. 개별 방법들의 장점들을 효과적으로 활용하기 위해, 판별 거리 척도를 이용하여 각 특징들의 분별력을 측정하여 분별력이 높은 특징들로만 복합 특징을 구성하여 얼굴 인식에 사용한다. Multi-PIE, Yale B, AR, yale database들에 대한 실험 결과, 제안한 방법은 모든 database에 대해 개별 조명 불변 특징 방법들보다 우수한 인식 성능을 보여 주었다.

• 센서학회지
• /
• 제10권6호
• /
• pp.286-294
• /
• 2001

본 연구에서는 자체 제작한 3-3형 복합압전체 초음파 트랜스듀서와 SOFM(Self Organizing Feature Map) 신경회로망을 이용한 수중 3차원 물체인식특성에 대해 연구하였다. 자체 제작한 3-3형 복합압전체 소자는 수중 초음파 트랜스듀서 재료로서의 요구조건을 비교적 잘 만족하였다. 자체 제작한 3-3형 복합압전체 트랜스듀서와 SOFM 신경회로망을 이용하여 얻어진 4종의 인식물체(정사각기둥, 직사각기둥, 원통, 정삼각기둥)에 대한 전체적인 수중 물체인식률은 학습데이터인 경우에는 100%, 시험데이터는 94.0%를 나타내었다. 이들 결과로부터 자체 제작한 3-3형 복합압전체 초음파 트랜스듀서는 수중 물체인식용 트랜스듀서로서 응용될 수 있음을 알 수 있었다.

• 대한의용생체공학회:의공학회지
• /
• 제31권6호
• /
• pp.481-486
• /
• 2010

In this paper, we suggest an advanced algorithm, to recognize pedestrian/non-pedestrian using differential haar-like feature, which applies Adaboost algorithm to make a strong classification from weak classifications. First, we extract two feature vectors: horizontal haar-like feature and vertical haar-like feature. For the next, we calculate the proposed feature vector using differential haar-like method. And then, a strong classification needs to be obtained from weak classifications for composite recognition method using the differential area of horizontal and vertical haar-like. In the proposed method, we use one feature vector and one strong classification for the first stage of recognition. Based on our experiment, the proposed algorithm shows higher recognition rate compared to the traditional method for the pedestrian and non-pedestrian.

• 대한산업공학회지
• /
• 제22권3호
• /
• pp.441-458
• /
• 1996

Many researches on feature recognition have been performed up to now, but the general solution for recognizing arbitrary features has not been developed. The most popular research area in feature recognition is automatic extraction of 2.5 dimensional features, because they are frequently used in manufacturing field. In this paper, a faster and more convenient 2.5 dimensional feature recognition algorithm is proposed using a new strategy which is quite different from the existing ones. The proposed algorithm takes process planning into consideration. The algorithm is implemented in C++. By applying the algorithm to practical complicate examples, we verify that the algorithm is working very well.

• Genomics & Informatics
• /
• 제17권4호
• /
• pp.39.1-39.20
• /
• 2019

Analyzing patterns in data points embedded in linear and non-linear feature spaces is considered as one of the common research problems among different research areas, for example: data mining, machine learning, pattern recognition, and multivariate analysis. In this paper, data points are heterogeneous sets of biosequences (composite data points). A composite data point is a set of ordinary data points (e.g., set of feature vectors). We theoretically extend the derivation of the largest generalized eigenvalue-based distance metric D_ij(γ₁) in any linear and non-linear feature spaces. We prove that D_ij(γ₁) is a metric under any linear and non-linear feature transformation function. We show the sufficiency and efficiency of using the decision rule $\bar{{\delta}}_{{\Xi}i}$(i.e., mean of D_ij(γ₁)) in classification of heterogeneous sets of biosequences compared with the decision rules min_𝚵iand median_𝚵i. We analyze the impact of linear and non-linear transformation functions on classifying/clustering collections of heterogeneous sets of biosequences. The impact of the length of a sequence in a heterogeneous sequence-set generated by simulation on the classification and clustering results in linear and non-linear feature spaces is empirically shown in this paper. We propose a new concept: the limiting dispersion map of the existing clusters in heterogeneous sets of biosequences embedded in linear and nonlinear feature spaces, which is based on the limiting distribution of nucleotide compositions estimated from real data sets. Finally, the empirical conclusions and the scientific evidences are deduced from the experiments to support the theoretical side stated in this paper.

• Journal of Electrical Engineering and Technology
• /
• 제12권1호
• /
• pp.363-372
• /
• 2017

This paper presents a novel face recognition algorithm based on the deep convolution neural network and key point detection jointed local binary pattern methodology to enhance the accuracy of face recognition. We firstly propose the modified face key feature point location detection method to enhance the traditional localization algorithm to better pre-process the original face images. We put forward the grey information and the color information with combination of a composite model of local information. Then, we optimize the multi-layer network structure deep learning algorithm using the Fisher criterion as reference to adjust the network structure more accurately. Furthermore, we modify the local binary pattern texture description operator and combine it with the neural network to overcome drawbacks that deep neural network could not learn to face image and the local characteristics. Simulation results demonstrate that the proposed algorithm obtains stronger robustness and feasibility compared with the other state-of-the-art algorithms. The proposed algorithm also provides the novel paradigm for the application of deep learning in the field of face recognition which sets the milestone for further research.