• ETRI Journal
• /
• v.31 no.4
• /
• pp.438-447
• /
• 2009

In this paper, a novel image projection technique for face recognition application is proposed which is based on linear discriminant analysis (LDA) combined with the relevance-weighted (RW) method. The projection is performed through 2-directional and 2-dimensional LDA, or $(2D)^2$LDA, which simultaneously works in row and column directions to solve the small sample size problem. Moreover, a weighted discriminant hyperplane is used in the between-class scatter matrix, and an RW method is used in the within-class scatter matrix to weigh the information to resolve confusable data in these classes. This technique is called the relevance-weighted $(2D)^2$LDA, or RW$(2D)^2$LDA, which is used for a more accurate discriminant decision than that produced by the conventional LDA or 2DLDA. The proposed technique has been successfully tested on four face databases. Experimental results indicate that the proposed RW$(2D)^2$LDA algorithm is more computationally efficient than the conventional algorithms because it has fewer features and faster times. It can also improve performance and has a maximum recognition rate of over 97%.

• Proceedings of the KIEE Conference
• /
• 2003.11c
• /
• pp.1006-1009
• /
• 2003

In this paper, we propose a technique of the recognition of partial face. Most of the research is concentrated on the recognition of whole face Since part of the face area in an image can be damaged or overlapped, face recognition based on partial face is required. PCA and LDA technique is applied to the recognition of partial face. Also, a new method to combine the results of the recognition of parts of the face.

• Journal of Digital Contents Society
• /
• v.19 no.1
• /
• pp.51-58
• /
• 2018

As the amount of users and data of NS explosively increased, research based on SNS Big data became active. In social mining, Latent Dirichlet Allocation(LDA), which is a typical topic model technique, is used to identify the similarity of each text from non-classified large-volume SNS text big data and to extract trends therefrom. However, LDA has the limitation that it is difficult to deduce a high-level topic due to the semantic sparsity of non-frequent word occurrence in the short sentence data. The BTM study improved the limitations of this LDA through a combination of two words. However, BTM also has a limitation that it is impossible to calculate the weight considering the relation with each subject because it is influenced more by the high frequency word among the combined words. In this paper, we propose a technique to improve the accuracy of existing BTM by reflecting semantic relation between words.

• Journal of KIISE:Software and Applications
• /
• v.32 no.8
• /
• pp.781-788
• /
• 2005

LDA (Linear Discriminant Analysis) is a data discrimination technique that seeks transformation to maximize the ratio of the between-class scatter and the within-class scatter While it has been successfully applied to several applications, it has two limitations, both concerning the underfitting problem. First, it fails to discriminate data with complex distributions since all data in each class are assumed to be distributed in the Gaussian manner; and second, it can lose class-wise information, since it produces only one transformation over the entire range of classes. We propose three extensions of LDA to overcome the above problems. The first extension overcomes the first problem by modeling the within-class scatter using a PCA mixture model that can represent more complex distribution. The second extension overcomes the second problem by taking different transformation for each class in order to provide class-wise features. The third extension combines these two modifications by representing each class in terms of the PCA mixture model and taking different transformation for each mixture component. It is shown that all our proposed extensions of LDA outperform LDA concerning classification errors for handwritten digit recognition and alphabet recognition.

• Journal of Korea Society of Industrial Information Systems
• /
• v.22 no.4
• /
• pp.19-28
• /
• 2017

Companies and Governments have Mainly used the Delphi Technique to Understand Research or Technology Trends. Because this Technique has the Disadvantage of Consuming a Large Amount of Time and Money, this Study Attempted to Understand and Predict Science and Technology Trends using the Topic Modeling Technique Latent Dirichlet Allocation (LDA). To this end, 20 Specific Artificial Intelligence (AI) Technologies were Extracted From the Abstracts of the US Patent Documents on AI. With Regard to the Extracted Specific Technologies, Core Technologies were Identified, and then these were Divided into Hot and Cold Technologies though a Trend Analysis on their Annual Proportions. Text/Word Searching, Computer Management, Programming Syntax, Network Administration, Multimedia, and Wireless Network Technology were Derived From Hot Technologies. These Technologies are Key Technologies that are Actively Studied in the Field of AI in Recent Years. The Methodology Suggested in this Study may be used to Analyze Trends, Derive Policies, or Predict Technical Demands in Various Fields such as Social Issues, Regional Innovation, and Management.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38C no.2
• /
• pp.208-212
• /
• 2013

Due to the singularity of the within-class scatter, linear discriminant analysis (LDA) becomes ill-posed for small sample size (SSS) problems. An extension of LDA, the null space-based LDA (NLDA) provides good discriminant performances for SSS problems. In this paper, by applying the Lagrange technique, the procedure of transforming the problem of finding the feature extractor of NLDA into a linear equation problem is derived.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.12C
• /
• pp.1162-1166
• /
• 2005

A new low dimensional feature representation technique is presented in this paper. Linear discriminant analysis is a popular feature extraction method. However, in the case of high dimensional data, the computational difficulty and the small sample size problem are often encountered. In order to solve these problems, we propose two dimensional direct LDA algorithm, which directly extracts the image scatter matrix from 2D image and uses Direct LDA algorithm for face recognition. The ORL face database is used to evaluate the performance of the proposed method. The experimental results indicate that the performance of the proposed method is superior to DLDA.

• Journal of Korea Multimedia Society
• /
• v.10 no.7
• /
• pp.923-933
• /
• 2007

In this paper, we present a method for performance enhancement of the marker detection system by using SVM(Support Vector Machine) and LDA(Linear Discriminant Analysis). It converts the input image to a binary image and extracts contours of objects in the binary image. After that, it approximates the contours to a list of line segments. It finds quadrangle by using geometrical features which are extracted from the approximated line segments. It normalizes the shape of extracted quadrangle into exact squares by using the warping technique and scale transformation. It extracts feature vectors from the square image by using principal component analysis. It then checks if the square image is a marker image or a non-marker image by using a SVM classifier. After that, it computes feature vectors by using LDA for the extracted marker images. And it calculates the distance between feature vector of input marker image and those of standard markers. Finally, it recognizes the marker by using minimum distance method. Experimental results show that the proposed method achieves enhancement of recognition rate with smaller feature vectors by using LDA and it can decrease false detection errors by using SVM.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.3
• /
• pp.367-371
• /
• 2006

Multilayer percepteon has the advantage of learning their optimal parameters and efficiency. However, MLP shows some drawbacks when dealing with high dimensional data within the input space. Also, it Is very difficult to find the optimal parameters when the input data are highly correlated such as large scale face dataset. In this paper, we propose a novel technique for face recognition based on LDA and local MLP. To resolve the main drawback of MLP, we calculate the reduced features by LDA in advance. And then, we construct a local MLP per group consisting of subset of facedatabase to find its optimal learning parameters rather than using whole faces. Finally, we designed the face recognition system combined with the local MLPs. From various experiments, we obtained better classification performance in comparison with the results produced by conventional methods such as PCA and LDA.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.7
• /
• pp.73-80
• /
• 2024

In this paper, we proposes a method to improve the accuracy of predicting the Korea Composite Stock Price Index (KOSPI) by combining topic modeling and Long Short-Term Memory (LSTM) neural networks. In this paper, we use the Latent Dirichlet Allocation (LDA) technique to extract ten major topics related to interest rate increases and decreases from financial news data. The extracted topics, along with historical KOSPI index data, are input into an LSTM model to predict the KOSPI index. The proposed model has the characteristic of predicting the KOSPI index by combining the time series prediction method by inputting the historical KOSPI index into the LSTM model and the topic modeling method by inputting news data. To verify the performance of the proposed model, this paper designs four models (LSTM_K model, LSTM_KNS model, LDA_K model, LDA_KNS model) based on the types of input data for the LSTM and presents the predictive performance of each model. The comparison of prediction performance results shows that the LSTM model (LDA_K model), which uses financial news topic data and historical KOSPI index data as inputs, recorded the lowest RMSE (Root Mean Square Error), demonstrating the best predictive performance.