• Journal of the Korea Society of Computer and Information
• /
• v.24 no.1
• /
• pp.33-39
• /
• 2019

In this paper, we propose deep convolutional neural network(CNN) with bottleneck structure which improves the performance of earthquake classification. In order to address all possible forms of earthquakes including micro-earthquakes and artificial-earthquakes as well as large earthquakes, we need a representation and classifier that can effectively discriminate seismic waveforms in adverse conditions. In particular, to robustly classify seismic waveforms even in low snr, a deep CNN with 1x1 convolution bottleneck structure is proposed in raw seismic waveforms. The representative experimental results show that the proposed method is effective for noisy seismic waveforms and outperforms the previous state-of-the art methods on domestic earthquake database.

• Korean Institute of Interior Design Journal
• /
• v.24 no.1
• /
• pp.207-215
• /
• 2015

It seems obvious that the spatial information of existing architectural heritage will be re-structured utilizing BIM technology. In the future to be able to implement such task, a new system of classification of spatial information, which fit to the structural nature of architectural heritage is necessary. This paper intend to suggest the conceptual model that can be the base of establishing new classification system for architectural heritage. For this study we reviewed researches related to classification system of architectural heritage (CS-AH) and BIM based architectural heritage (BIM-AH), first. As a result, we found that CS-AH is focused on building elevation and type, and BIM-AH is biased on the Library and Parametric Modeling. Second, we figured out a relationship between the CS-AH and BIM-AH. From this analysis, we found that BIM-AH is biased on Library and Parametric since the building elevation and type was focused on CS-AH. This review suggests a potential of the 3D CS-AH to expand the range of research for BIM-AH. At last, we suggest the three concept of classification are: 1)horizontality-accumulation relationship, 2)structure-infill relationship, 3)segment-member relationship. These three concept, together as one system of classification, could provide useful framework of new classification system of spatial information for architectural heritage.

• Convergence Security Journal
• /
• v.18 no.3
• /
• pp.53-59
• /
• 2018

In this paper, we propose a hierarchical image classification scheme for efficient object image classification. In the non-hierarchical image classification, which classifies the existing whole images at one time, it showed that objects with relatively similar shapes are not recognized efficiently. Therefore, in this paper, we introduce the image classification method in the hierarchical structure which attempts to classify object images hierarchically. Also, we introduce to the efficient class structure and algorithms considering the scalability that can occur when a deep learning image classification is applied to an actual system. Such a scheme makes it possible to classify images with a higher degree of confidence in object images having relatively similar shapes.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.1
• /
• pp.9-23
• /
• 2019

In this paper, we propose a layer structure of a pest image classifier model using CNN (Convolutional Neural Network) and background removal image processing algorithm for improving classification accuracy in order to build a smart monitoring system for pine wilt pest control. In this study, we have constructed and trained a CNN classifier model by collecting image data of pine wilt pest mediators, and experimented to verify the classification accuracy of the model and the effect of the proposed classification algorithm. Experimental results showed that the proposed method successfully detected and preprocessed the region of the object accurately for all the test images, resulting in showing classification accuracy of about 98.91%. This study shows that the layer structure of the proposed CNN classifier model classified the targeted pest image effectively in various environments. In the field test using the Smart Trap for capturing the pine wilt pest mediators, the proposed classification algorithm is effective in the real environment, showing a classification accuracy of 88.25%, which is improved by about 8.12% according to whether the image cropping preprocessing is performed. Ultimately, we will proceed with procedures to apply the techniques and verify the functionality to field tests on various sites.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.767-772
• /
• 2005

The Internet is a packet switched network which offers best-effort service, but current IP network provide enhanced services such Quality of Services, Virtual Private Network (VPN) services, Distribute Firewall and IP Security Gateways. All such services need packet classification for determining the flow. The problem is performing scalable packet classification at wire speeds even as rule databases increase in size. Therefore, this research offer packet classification algorithm that increase classifier performance when working with enlarge rules database by rearrange rule structure into Bitmap Intersection Lookup (BIL) tables. It will use packet's header field for looking up BIL tables and take the result with intersection operation by logical AND. This approach will use simple algorithm and rule structure, it make classifier have high search speed and fast updates.

• Proceedings of the KIEE Conference
• /
• 1992.07a
• /
• pp.449-452
• /
• 1992

Before performing an adaptive stereo matching using satellite images, classification is needed as a preprocessing step. This paper describes that classification of three land cover types : river, mountain, and agricultural fields. We proposed that classification algorithm using pyramid structure and texture features. Results of applying the proposed algorithm to satellite image improved classification accuracy.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.45 no.6
• /
• pp.949-964
• /
• 2021

Accumulation of high-quality data is crucial for AI learning. The goal of using AI in fashion service is to propose of a creative, personalized solution that is close to the know-how of a human operator. These customized solutions require an understanding of fashion products and emotions. Therefore, it is necessary to accumulate data on the attributes of fashion products and fashion emotion. The first step for accumulating fashion data is to standardize the attribute with coherent system. The purpose of this study is to propose a fashion emotional classification system. For this, images of fashion products were collected, and metadata was obtained by allowing consumers to describe their emotions about fashion images freely. An emotional classification system with a hierarchical structure, was then constructed by performing frequency and CONCOR analyses on metadata. A final classification system was proposed by supplementing attribute values with reference to findings from previous studies and SNS data.

• Journal of information and communication convergence engineering
• /
• v.21 no.1
• /
• pp.9-16
• /
• 2023

The natural language on social network platforms has a certain front-to-back dependency in structure, and the direct conversion of Chinese text into a vector makes the dimensionality very high, thereby resulting in the low accuracy of existing text classification methods. To this end, this study establishes a deep learning model that combines a big data ultra-deep convolutional neural network (UDCNN) and long short-term memory network (LSTM). The deep structure of UDCNN is used to extract the features of text vector classification. The LSTM stores historical information to extract the context dependency of long texts, and word embedding is introduced to convert the text into low-dimensional vectors. Experiments are conducted on the social network platforms Sogou corpus and the University HowNet Chinese corpus. The research results show that compared with CNN + rand, LSTM, and other models, the neural network deep learning hybrid model can effectively improve the accuracy of text classification.

• Journal of the Korean Society for information Management
• /
• v.34 no.4
• /
• pp.33-57
• /
• 2017

This study aims to introduce an emerging prescriptive analytics method and suggest its efficient application to a category-based service system. Prescriptive analytics method provides the whole process of analysis and available alternatives as well as the results of analysis. To simulate the process of optimization, large scale journal articles have been collected and categorized by classification scheme. In the process of applying the concept of prescriptive analytics to a real system, we have fused a dynamic automatic-categorization method for large scale documents and intellectual structure analysis method for scholarly subject fields. The test result shows that some optimized scenarios can be generated efficiently and utilized effectively for reorganizing the classification-based service system.

• The Journal of Society for e-Business Studies
• /
• v.25 no.3
• /
• pp.77-93
• /
• 2020

The literature has reported that hierarchical classification methods generally outperform the flat classification methods for a multi-class document classification problem. Unlike the literature that has constructed a class hierarchy, this paper evaluates the performance of hierarchical and flat classification methods under a situation where the class hierarchy is predefined. We conducted numerical evaluations for two data sets; research papers on climate change adaptation technologies in water sector and 20NewsGroup open data set. The evaluation results show that the hierarchical classification method outperforms the flat classification methods under a certain condition, which differs from the literature. The performance of hierarchical classification method over flat classification method depends on class similarities at levels in the class structure. More importantly, the hierarchical classification method works better when the upper level similarity is less that the lower level similarity.