• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1995.04a
• /
• pp.587-599
• /
• 1995

This paper proposes a method for pattern recognition using spectral analysis and a hybrid genetic-neural-fuzzy networks. The feature vectors using spectral analysis on contour sequences of 2-D images are extracted, and the vectors are not effected by translation, rotation and scale variance. A combined model using the advantages of conventional method is proposed, those are supervised learning BP, global searching genetic algorithm, and unsupervised learning fuzzy c-method. The proposed method is applied to 10 aircraft recognition to confirm the performance of the method. The experimental results show that the proposed method is better accuracy than conventional method using BP or fuzzy c-method, and learning speed is enhanced.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.4
• /
• pp.2093-2108
• /
• 2017

In this paper, we describe a novel method for recognizing human actions from different views via view knowledge transfer. Our approach is characterized by two aspects: 1) We propose a unsupervised topic transfer model (TTM) to model two view-dependent vocabularies, where the original bag of visual words (BoVW) representation can be transferred into a bag of topics (BoT) representation. The higher-level BoT features, which can be shared across views, can connect action models for different views. 2) Our features make it possible to obtain a discriminative model of action under one view and categorize actions in another view. We tested our approach on the IXMAS data set, and the results are promising, given such a simple approach. In addition, we also demonstrate a supervised topic transfer model (STTM), which can combine transfer feature learning and discriminative classifier learning into one framework.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.24 no.4
• /
• pp.382-391
• /
• 2021

Unsupervised domain adaptation often gives impressive solutions to handle domain shift of data. Most of current approaches assume that unlabeled target data to train is abundant. This assumption is not always true in practices. To tackle this issue, we propose a general solution to solve the domain gap minimization problem without any target data. Our method consists of two regularization steps. The first step is a pixel regularization by arbitrary style transfer. Recently, some methods bring style transfer algorithms to domain adaptation and domain generalization process. They use style transfer algorithms to remove texture bias in source domain data. We also use style transfer algorithms for removing texture bias, but our method depends on neither domain adaptation nor domain generalization paradigm. The second regularization step is a feature regularization by feature alignment. Adding a feature alignment loss term to the model loss, the model learns domain invariant representation more efficiently. We evaluate our regularization methods from several experiments both on small dataset and large dataset. From the experiments, we show that our model can learn domain invariant representation as much as unsupervised domain adaptation methods.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.316-318
• /
• 2001

본 논문에서는 자기조직화 형상지도(Self-organizing Feature Maps)를 사용하여 움직이는 물체에 대해 움직임의 특성을 자동으로 분석하였다. Kohonen Network는 자기조직을 형성하는 unsupervised learning 알고리즘으로서, 이 논문에서는 생태계에서의 데이터를 Patternizing하고, Clustering 하는데 사용한다. 본 논문에서 Kohonen 신경망의 학습에 사용한 데이터는 CCD 카메라로 물고기의 움직임을 추적한 좌표 데이터이며, diazinon 0.1 ppm을 처리한 물고기 점 데이터와 처리하지 않은 점 데이터를 각각 낮.밤 약 10시간동안 수집하여, \circled1처리전 낮 데이터 \circled2처리전 밤 데이터 \circled3처리전 낮 데이터 \circled4처리후 밤 데이터 각각 4개의 group으로 분류한 후, Kohonen Network을 사용하여 물고기의 행동 차이를 분석하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10c
• /
• pp.739-742
• /
• 2002

This paper addresses a new statistical method, IFSAcycle, which is an unsupervised learning method of analyzing cell cycle-related gene expression data. The IFSAcycle is based on the independent feature subspace analysis (IFAS) [3], which generalizes the independent component analysis (ICA). Experimental results show the usefulness of IFAS: (1) the ability of assigning genes to multiple coexpression pattern groups; (2) the capability of clustering key genes that determine each critical point of cell cycle.

• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.327-336
• /
• 2002

Texture analysis is an important technique in many image understanding areas, such as perception of surface, object, shape and depth. But the previous works are intend to the issue of only texture segment, that is not capable of acquiring recognition information. No unsupervised method is basased on the recognition of texture in image. we propose a novel approach for efficient texture image analysis that uses unsupervised learning schemes for the texture recognition. The self-organization neural network for multiple texture image identification is based on block-based clustering and merging. The texture features used are the angle and magnitude in orientation-field that might be different from the sample textures. In order to show the performance of the proposed system, After we have attempted to build a various texture images. The final segmentation is achieved by using efficient edge detection algorithm applying to block-based dilation. The experimental results show that the performance of the system Is very successful.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.6
• /
• pp.2511-2520
• /
• 2018

In this study, Multi-Radial Basis Function Support Vector Machine (Multi-RBF SVM) classifier is introduced based on a composite kernel function. In the proposed multi-RBF support vector machine classifier, the input space is divided into several local subsets considered for extremely nonlinear classification tasks. Each local subset is expressed as nonlinear classification subspace and mapped into feature space by using kernel function. The composite kernel function employs the dual RBF structure. By capturing the nonlinear distribution knowledge of local subsets, the training data is mapped into higher feature space, then Multi-SVM classifier is realized by using the composite kernel function through optimization procedure similar to conventional SVM classifier. The original training data set is partitioned by using some unsupervised learning methods such as clustering methods. In this study, three types of clustering method are considered such as Affinity propagation (AP), Hard C-Mean (HCM) and Iterative Self-Organizing Data Analysis Technique Algorithm (ISODATA). Experimental results on benchmark machine learning datasets show that the proposed method improves the classification performance efficiently.

• Annual Conference on Human and Language Technology
• /
• 1995.10a
• /
• pp.39-42
• /
• 1995

문장 안에서 한 단어가 가지는 올바른 의미를 얻기 위해 모듈화된 신경망을 이용하였다. 앞부분에 놓인 신경망은 코호넨 신경망으로 사용자의 지도가 개입되지 않은 상태로 자율학습(Unsupervised learning)이 이루어지고, 뒤에 놓인 신경망은 앞에서 결과로 얻은 2차원의 자기 조직화 형상지도(Self-organizing feature map)를 바탕으로 역전파 신경망을 이용한 지도학습(Supervised learning)을 하게 하였다. 입력 자료는 구문분석된 문장의 조사 정보를 활용하여 입력 위치를 정해준 명사의 의미표지와 동사의 의미표지를 사용하였다. 중의성이 있는 단어를 가지는 문장은 중의성의 가지수 만큼 테스트 입력 자료가 되어 신경망을 통과하여 의미를 결정하도록 한다.

• ETRI Journal
• /
• v.46 no.3
• /
• pp.513-525
• /
• 2024

Deep neural networks trained on labeled medical data face major challenges owing to the economic costs of data acquisition through expensive medical imaging devices, expert labor for data annotation, and large datasets to achieve optimal model performance. The heterogeneity of diseases, such as Alzheimer's disease, further complicates deep learning because the test cases may substantially differ from the training data, possibly increasing the rate of false positives. We propose a reconstruction-based self-supervised anomaly detection model to overcome these challenges. It has a dual-subnetwork encoder that enhances feature encoding augmented by skip connections to the decoder for improving the gradient flow. The novel encoder captures local and global features to improve image reconstruction. In addition, we introduce an entropy-based image conversion method. Extensive evaluations show that the proposed model outperforms benchmark models in anomaly detection and classification using an encoder. The supervised and unsupervised models show improved performances when trained with data preprocessed using the proposed image conversion method.

• Science of Emotion and Sensibility
• /
• v.20 no.2
• /
• pp.161-170
• /
• 2017

Emotion affects many parts of human life such as learning ability, behavior and judgment. It is important to understand human nature. Emotion can only be inferred from facial expressions or gestures, what it actually is. In particular, emotion is difficult to classify not only because individuals feel differently about emotion but also because visually induced emotion does not sustain during whole testing period. To solve the problem, we acquired bio-signals and extracted features from those signals, which offer objective information about emotion stimulus. The emotion pattern classifier was composed of unsupervised learning algorithm with hidden nodes and feature vectors. Restricted Boltzmann machine (RBM) based on probability estimation was used in the unsupervised learning and maps emotion features to transformed dimensions. The emotion was characterized by non-linear classifiers with hidden nodes of a multi layer neural network, named deep belief network (DBN). The accuracy of DBN (about 94 %) was better than that of back-propagation neural network (about 40 %). The DBN showed good performance as the emotion pattern classifier.