• Journal of Institute of Control, Robotics and Systems
• /
• v.13 no.8
• /
• pp.754-759
• /
• 2007

In this paper, we proposed the Bi-Modal Sensor Fusion Algorithm which is the emotional recognition method that be able to classify 4 emotions (Happy, Sad, Angry, Surprise) by using facial image and speech signal together. We extract the feature vectors from speech signal using acoustic feature without language feature and classify emotional pattern using Neural-Network. We also make the feature selection of mouth, eyes and eyebrows from facial image. and extracted feature vectors that apply to Principal Component Analysis(PCA) remakes low dimension feature vector. So we proposed method to fused into result value of emotion recognition by using facial image and speech.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.703-708
• /
• 2009

The present paper describes the application of an improved impression transfer vector method (Sakurai et al., 2007) to transform the three basic dimensions (Evaluation, Activity, and Potency) of higher-order impression. First, a set of shapes and surface textures of faces was represented by multi-dimensional vectors. Second, the variation among faces was coded in reduced parameters derived by applying principal component analysis. Third, a facial attribute along a given impression dimension was analyzed to select discriminative parameters from among principal components with higher sensitivity to impressions, and obtain an impression transfer vector. Finally, the parametric coordinates were changed by adding or subtracting the impression transfer vector and the image was manipulated so that its facial appearance clearly exhibits the transformed impression. A psychological rating experiment confirmed that the impression transfer vector modulated three dimensions of higher-order impression. We discussed the versatility of the impression transfer vector method.

• ETRI Journal
• /
• v.40 no.5
• /
• pp.634-642
• /
• 2018

Projection spectral analysis is investigated and refined in this paper, in order to unify principal component analysis and independent component analysis. Singular value decomposition and spectral theorems are applied to nonsymmetric correlation or covariance matrices with multiplicities or singularities, where projections and nilpotents are obtained. Therefore, the suggested approach not only utilizes a sum-product of orthogonal projection operators and real distinct eigenvalues for squared singular values, but also reduces the dimension of correlation or covariance if there are multiple zero eigenvalues. Moreover, incremental learning strategies of projection spectral analysis are also suggested to improve the performance.

• Communications for Statistical Applications and Methods
• /
• v.11 no.3
• /
• pp.539-551
• /
• 2004

Multiple responses are often observed in many application fields, such as customer's time-of-day pattern for using internet. Some decision trees for multiple responses have been constructed by many researchers. However, if the response is a high-dimensional vector that can be thought of as a discretized function, then fitting a multivariate decision tree may be unsuccessful. Yu and Lambert (1999) suggested spline tree and principal component tree to analyze high dimensional response vector by using dimension reduction techniques. In this paper, we shall propose factor tree which would be more interpretable and competitive. Furthermore, using Korean internet company data, we will analyze time-of-day patterns for internet user.

• International Journal of Contents
• /
• v.14 no.4
• /
• pp.24-29
• /
• 2018

EEG-based brain-computer interfaces has focused on explicitly expressed intentions to assist physically impaired patients. For EEG-based-computer interfaces to function effectively, it should be able to understand users' implicit information. Since it is hard to gather EEG signals of human brains, we do not have enough training data which are essential for proper classification performance of implicit intention. In this paper, we improve the subject independent classification of implicit intention through the generation of additional training data. In the first stage, we perform the PCA (principal component analysis) of training data in a bid to remove redundant components in the components within the input data. After the dimension reduction by PCA, we train ICA (independent component analysis) network whose outputs are statistically independent. We can get additional training data by adding Gaussian noises to ICA outputs and projecting them to input data domain. Through simulations with EEG data provided by CNSL, KAIST, we improve the classification performance from 65.05% to 66.69% with Gamma components. The proposed sample generation method can be applied to any machine learning problem with fewer samples.

• The Korean Journal of Applied Statistics
• /
• v.28 no.3
• /
• pp.385-392
• /
• 2015

Principal component analysis (PCA), a dimension-reduction technique, is usually implemented after the variables are standardized when the measurement unit of variables are different. To standardize a variable we divide it by its standard deviation. But there is another way to transform a variable to be independent of its measurement unit. It is to divide it by its mean rather than standard deviation. Implementing PCA on standardized variables is equivalent to implementing PCA with a correlation matrix of original variables. Similarly, implementing PCA on the transformed variables divided by their means is equivalent to implementing PCA with a matrix related to the coefficients of variation of the original variables. We explain why we need to implement PCA on the variables transformed by their means.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2000.12a
• /
• pp.181-184
• /
• 2000

The new method of feature extraction is proposed, considering the statistic feature of human voice, unlike the conventional methods of voice extraction. PCA(principal Component Analysis) is applied to this new method. PCA removes the repeating of data after finding the axis direction which has the greatest variance in input dimension. Then the new method is applied to real voice recognition to assess performance. When results of the number recognition in this paper and the conventional Mel-Cepstrum of voice feature parameter are compared, there is 0.5% difference of recognition rate. Better recognition rate is expected than word or sentence recognition in that less convergence time than the conventional method in extracting voice feature. Also, better recognition tate is expected when the optimum vector is used by statistic feature of data.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.09a
• /
• pp.372-375
• /
• 2003

The incidence of blindness resulting from diabetic retinopathy has significantly increased despite the intervention of insulin to control diabetes mellitus. Early signs are microaneurysms, exudates, intraretinal hemorrhages, cotton wool patches, microvascular abnormalities, and venous beading. Advanced stages include neovascularization, fibrous formations, preretinal and vitreous microhemorrhages, and retinal detachment. Microaneurysm count is important because it is an indicator of retinopathy progression. The purpose of this paper is to apply data mining to detect diabetic retinopathy patterns in routine fundus fluorescein angiography. Early symptoms are of principal interest and therefore the emphasis is on detecting microaneurysms rather than vessel tortuosity. The analysis does not involve image-recognition algorithms. Instead, mathematical filtering isolates microaneurysms, microhemorrhages, and exudates as objects of disconnected sets. A neural network is trained on their distribution to return fractal dimension. Hausdorff and box counting dimensions grade progression of the disease. The field is acquired on fluorescein angiography with resolution superior to color ophthalmoscopy, or on patterns produced by physical or mathematical simulations that model viscous fingering of water with additives percolated through porous media. A mathematical filter and neural network perform the screening process thereby eliminating the time consuming operation of determining fractal set dimension in every case.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.3E
• /
• pp.89-94
• /
• 2006

For audio indexing and targeted search of specific audio or corresponding visual contents, the MPEG-7 standard has adopted a sound classification framework, in which dimension-reduced Audio Spectrum Projection (ASP) features are used to train continuous hidden Markov models (HMMs) for classification of various sounds. The MPEG-7 employs Principal Component Analysis (PCA) or Independent Component Analysis (ICA) for the dimensional reduction. Other well-established techniques include Non-negative Matrix Factorization (NMF), Linear Discriminant Analysis (LDA) and Discrete Cosine Transformation (DCT). In this paper we compare the performance of different dimensional reduction methods with Gaussian mixture models (GMMs) and HMMs in the classifying video sound clips.

• The Mathematical Education
• /
• v.46 no.3
• /
• pp.263-272
• /
• 2007

Mathematics Teaching Efficacy Beliefs Inventory (MTEBI) was translated into Korean and conducted among Korean pre-service mathematics teachers. The Korean-translated MTEBI consists of two subscales with 16 items. Personal Mathematics Teaching Efficacy (PMTE) subscale has 10 items and Mathematics Teaching Outcome Expectancy (MTOE) subscale has 6 items. The purpose of this study is to investigate the internal reliability and the construct validity of the Korean-translated MTEBI. The Cronbach alpha coefficient of Korean-translated MTEBI and its two subscales are respectively .87, .83, and .74 which imply that the instrument is reliable. The construct validity was achieved by performing factor analysis. Principal component solution with varimax rotation for the Korean-translated MTEBI was used in factor analysis and thus the best fit simple structure was obtained by two factors which correspond to the self-efficacy dimension and the outcome expectancy dimension of Bandura's self-efficacy theory.