• Journal of the Korea Society of Computer and Information
• /
• v.28 no.4
• /
• pp.31-39
• /
• 2023

In this paper, we propose an Internal/External Knowledge Distillation (IEKD), which utilizes both external correlations between feature maps of heterogeneous models and internal correlations between feature maps of the same model for transferring knowledge from a teacher model to a student model. To achieve this, we transform feature maps into a sequence format and extract new feature maps suitable for knowledge distillation by considering internal and external correlations through a transformer. We can learn both internal and external correlations by distilling the extracted feature maps and improve the accuracy of the student model by utilizing the extracted feature maps with feature matching. To demonstrate the effectiveness of our proposed knowledge distillation method, we achieved 76.23% Top-1 image classification accuracy on the CIFAR-100 dataset with the "ResNet-32×4/VGG-8" teacher and student combination and outperformed the state-of-the-art KD methods.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.7 no.4
• /
• pp.649-655
• /
• 2006

This paper proposes a HMM-based recognition method using DMSVQ(Dynamic Multi-Section Vector Quantization) codebook by DMS model and double spectral feature, as a method on the speech recognition of speaker-independent. LPC cepstrum parameter is used as a instantaneous spectral feature and LPC cepstrum's regression coefficient is used as a dynamic spectral feature These two spectral features are quantized as each VQ codebook. HMM using DMS model is modeled by receiving instantaneous spectral feature and dynamic spectral feature by input. Other experiments to compare with the results of recognition experiments using proposed method are implemented by the various conventional recognition methods under the equivalent environment of data and conditions. Through the experiment results, it is proved that the proposed method in this paper is superior to the conventional recognition methods.

• Korean Journal of Remote Sensing
• /
• v.23 no.5
• /
• pp.401-408
• /
• 2007

Animated anaglyph is the most cost-effective method for 3D stereo visualization of virtual or actual 3D geo-based data model. Unlike 3D anaglyph scene generation using paired epipolar images, the main data sets of this study is the multi-typed 3D feature model containing 3D shaped objects, DEM and satellite imagery. For this purpose, a prototype implementation for 3D animated anaglyph using OpenGL API is carried out, and virtual 3D feature modeling is performed to demonstrate the applicability of this anaglyph approach. Although 3D features are not real objects in this stage, these can be substituted with actual 3D feature model with full texture images along all facades. Currently, it is regarded as the special viewing effect within 3D GIS application domains, because just stereo 3D viewing is a part of lots of GIS functionalities or remote sensing image processing modules. Animated anaglyph process can be linked with real-time manipulation process of 3D feature model and its database attributes in real world problem. As well, this approach of feature-based 3D animated anaglyph scheme is a bridging technology to further image-based 3D animated anaglyph rendering system, portable mobile 3D stereo viewing system or auto-stereo viewing system without glasses for multi-viewers.

• Geomechanics and Engineering
• /
• v.25 no.1
• /
• pp.1-16
• /
• 2021

Machine learning models have been widely used for landslide susceptibility assessment (LSA) in recent years. The large number of inputs or conditioning factors for these models, however, can reduce the computation efficiency and increase the difficulty in collecting data. Feature selection is a good tool to address this problem by selecting the most important features among all factors to reduce the size of the input variables. However, two important questions need to be solved: (1) how do feature selection methods affect the performance of machine learning models? and (2) which feature selection method is the most suitable for a given machine learning model? This paper aims to address these two questions by comparing the predictive performance of 13 feature selection-based machine learning (FS-ML) models and 5 ordinary machine learning models on LSA. First, five commonly used machine learning models (i.e., logistic regression, support vector machine, artificial neural network, Gaussian process and random forest) and six typical feature selection methods in the literature are adopted to constitute the proposed models. Then, fifteen conditioning factors are chosen as input variables and 1,017 landslides are used as recorded data. Next, feature selection methods are used to obtain the importance of the conditioning factors to create feature subsets, based on which 13 FS-ML models are constructed. For each of the machine learning models, a best optimized FS-ML model is selected according to the area under curve value. Finally, five optimal FS-ML models are obtained and applied to the LSA of the studied area. The predictive abilities of the FS-ML models on LSA are verified and compared through the receive operating characteristic curve and statistical indicators such as sensitivity, specificity and accuracy. The results showed that different feature selection methods have different effects on the performance of LSA machine learning models. FS-ML models generally outperform the ordinary machine learning models. The best FS-ML model is the recursive feature elimination (RFE) optimized RF, and RFE is an optimal method for feature selection.

• MALSORI
• /
• no.67
• /
• pp.167-180
• /
• 2008

This paper describes a novel spectral envelope conversion method based on Gaussian mixture model (GMM). The core of this paper is rearranging source feature vectors in input space to the transformed feature vectors in feature space for the better modeling of GMM of source and target features. The quality of statistical modeling is dependent on the distribution and the dimension of data. The proposed method transforms both of the distribution and dimension of data and gives us the chance to model the same data with different configuration. Because the converted feature vectors should be on the input space, only source feature vectors are rearranged in the feature space and target feature vectors remain unchanged for the joint pdf of source and target features using KPCA. The experimental result shows that the proposed method outperforms the conventional GMM-based conversion method in various training environment.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.52 no.5
• /
• pp.277-282
• /
• 2003

In this paper we propose Using Higher Order Neuron on the Supervised Learning Machine of the Kohonen Feature Map. The architecture of proposed model adopts the higher order neuron in the input layer of Kohonen Feature Map as a Supervised Learning Machine. It is able to estimate boundary on input pattern space because or the higher order neuron. However, it suffers from a problem that the number of neuron weight increases because of the higher order neuron in the input layer. In this time, we solved this problem by placing the second order neuron among the higher order neuron. The feature of the higher order neuron can be mapped similar inputs on the Kohonen Feature Map. It also is the network with topological mapping. We have simulated the proposed model in respect of the recognition rate by XOR problem, discrimination of 20 alphabet patterns, Mirror Symmetry problem, and numerical letters Pattern Problem.

• Journal of Ship and Ocean Technology
• /
• v.9 no.3
• /
• pp.33-42
• /
• 2005

As the networks connect the world, enterprises tend to move manufacturing activities into virtual spaces. Since different software applications use different data terminology, it becomes a problem to interoperate, interchange, and manage electronic data among heterogeneous systems. It is said that approximately one billion dollar has been being spent yearly in USA for product data exchange and interoperability. As commercial CAD systems have brought in the concept of design feature for the sake of interoperability, terminologies of design features need to be harmonized. In order to define design feature terminology for integration, knowledge about feature definitions of different CAD systems should be considered. STEP standard have attempted to solve this problem, but it defines only syntactic data representation so that semantic data integration is not possible. This paper proposes a methodology for integrating modeling features of CAD systems. We utilize the ontology concept to build a data model of design features which can be a semantic standard of feature definitions of CAD systems. Using feature ontology, we implement an integrated virtual database and a simple system which searches and edits design features in a semantic way.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.13 no.3
• /
• pp.30-36
• /
• 2004

The exchange of model design data among heterogeneous CAD systems is very difficult because each CAD system has different data structures suitable for its own functions. STEP represents product information in a common computer-interpretable form that is required to remain complete and consistent when the product information is needed to be exchanged among different computer systems. However, STEP has complex architecture to represent point, line, curve and vectors of element. Moreover it can't represent geometry data of feature based models. In this study, a structure of XML document that represents geometry data of feature based models as neutral format has been developed. To use the developed XML document, a converter also has been developed to exchange modules so that it can exchange feature based data models among heterogeneous CAD systems. Developed XML document and Converter have been applied to commercial CAD systems.

• Proceedings of the Korean Society of Machine Tool Engineers Conference
• /
• 2003.04a
• /
• pp.434-440
• /
• 2003

The exchange of model design date among heterogeneous CAD systems is a difficult task because each system has different data structures suitable for its own functions. STEP has been able to represent product information as a common computer-interpretable form that is required to remain complete and consistent when the product informant is needed to be exchanged among different computer system. However, STEP has difficult architecture in is representing point, line, curve and vectors of element, more over it can't represent geometry data of feature based models. In this study, a structure of XML document that represents geometry data of feature based models as neutral format has been developed. To use the developed XML document, a Converter has also been developed to exchange modules so that it can exchange feature based data models among heterogeneous CAD systems. Aa for evaluation of the developed XML document and Converter, Solidworks and SolidEdge are selected.

• Proceedings of the KSRS Conference
• /
• 2008.10a
• /
• pp.433-436
• /
• 2008

A demand on the spatial data management has been rapidly increased with the introduction and diffusion process of ITS, Telematics, and Wireless Sensor Network, and many different people use the digital map that offers various thematic spatial data. Spatial data for digital map can manage to tile-based and feature-based data. The existing tile-based digital map management systems have difficult problems of data construction, history management, and updating based on a spatial object. In order to solve these problems, this paper proposed the data model for the feature-based digital map management system that is designed for feature-based seamless map, history management, real-time updating of spatial data, and analyzed the validity and utility of the proposed model.