• 한국조사연구학회:학술대회논문집
• /
• 한국조사연구학회 2001년도 추계학술대회 발표논문집
• /
• pp.33-45
• /
• 2001

많은 통계조사에서 사용되고 있는 산업/직업분류코드가 기존에는 사람에 의해 수동으로 부호화되어 왔는데, 이러한 작업은 시간과 인력면에서 고비용을 요구할 뿐 아니라 개인별 시각, 이해도의 차이 등으로 정확성에 많은 문제가 제기되어 왔다. 본 논문에서는 이러한 수동코딩 작업의 문제점을 해결하기 위하여 자동코딩 시스템을 개발, 이를 인구 주택총조사와 사업체기초통계조사에 시험 적용하여 본 바를 바탕으로 향후 자동시스템으로의 전환 방향 등을 제시하고 있다.

• 대한건축학회연합논문집
• /
• 제20권6호
• /
• pp.89-96
• /
• 2018

The purpose of this study is to categorize Commercial area by identifying characteristics of blocks and coding them in order to segment use zoning in Commercial area. The study was conducted as follows. Data from building register, cadastral map, statistics annual report are utilized to identify the physical environment of the block. four types used as code under the physical environment classification code which are classification code of physical environment, detail usage, volume ratio, and height type are set, and combine the classification codes sorted by the four types of code. Through the physical environment classification codes, there are currently 37 different block characteristics of the Old downtown Commercial area. Diversity is not reflected. There are only Central commercial area of regulations in Old downtown commercial areas that are uniformly managed. For the renewal, management and development that can occur in the near future, it is necessary to segment of use district in the commercial area. Consider the current situation and future development direction for the management of sustainable commercial areas. Management is required using physical environment classification codes. It is meaningful that it can be maintained, managed and developed in accordance with the characteristics of each block.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1996년도 하계학술대회 논문집 B
• /
• pp.1353-1355
• /
• 1996

This paper presents classification of high impedance fault pattern using linear prediction coefficients. A feature of neutral phase current is extracted by the linear predictive coding. This feature is classified into faults by a multilayer perceptron neural network. Neural network successfully classifies test data into three faults and one normal state.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권5호
• /
• pp.2590-2606
• /
• 2017

Generally, research of classical image classification algorithms assume that training data and testing data are derived from the same domain with the same distribution. Unfortunately, in practical applications, this assumption is rarely met. Aiming at the problem, a domain adaption image classification approach based on multi-sparse representation is proposed in this paper. The existences of intermediate domains are hypothesized between the source and target domains. And each intermediate subspace is modeled through online dictionary learning with target data updating. On the one hand, the reconstruction error of the target data is guaranteed, on the other, the transition from the source domain to the target domain is as smooth as possible. An augmented feature representation produced by invariant sparse codes across the source, intermediate and target domain dictionaries is employed for across domain recognition. Experimental results verify the effectiveness of the proposed algorithm.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2018년도 추계학술대회
• /
• pp.11-14
• /
• 2018

The aim of this study is to rectify the misclassified image features and enhance the performance of image classification tasks by incorporating a channel- coding technique, widely used in telecommunication. Specifically, the proposed algorithm employs the error - correcting mechanism of convolutional coding combined with the convolutional neural networks (CNNs) that are the state - of- the- arts image classifier s. We develop an encoder and a decoder to employ the error - correcting capability of the convolutional coding. In the encoder, the label values of the image data are converted to convolutional codes that are used as target outputs of the CNN, and the network is trained to minimize the Euclidean distance between the target output codes and the actual output codes. In order to correct misclassified features, the outputs of the network are decoded through the trellis structure with Viterbi algorithm before determining the final prediction. This paper demonstrates that the proposed architecture advances the performance of the neural networks compared to the traditional one- hot encoding method.

• 한국통신학회논문지
• /
• 제19권11호
• /
• pp.2264-2275
• /
• 1994

본 논문에서는 블록 분류와 코딩과 함께 신경회로망을 이용한 영상압축을 보였다. 오차 역전파 알고리즘으로 학습되는 다층구조 신경회로망은 정규화된 영상데이타를 감소된 공간 중복성을 가지는 은닉층의 값으로 변환하는데 사용된다. 기본적으로 영상압축은 입력층과 출력층의 뉴런보다 적은 수의 은닉층 뉴런에 의해 얻어진다. 여기에 시각체계의 민감도에 따른 영상블럭 복잡성에 따라 적응적으로 압축되므로 블록을 분류한다. 또한 은닉뉴런의 양자화된 값은 효과적인 전송을 위해 entropy coding을 이용한 경우 화질의 큰 저하없이 약 25:1의 압축률을 얻었다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 The 6th International Symposium of East Asian Resources Recycling Technology
• /
• pp.693-698
• /
• 2001

This paper establishes a coding method system including the liberation and size distribution of recycling materials in the shredder operation. Every particle in the shredded product becomes a code number using the liberation model and size distribution equation transforming of weight percentage into particles number percentage. One set of database can be obtained after all particles have been coded. This database is suitable for the size reduction operation in the process simulation of waste recycling. Coupling with the developed air classification, sizing and separating operations, the whole process simulation will be completely established for diversified application. A typical simulation for the rolling cutting shredder product of waste TV had been demonstrated under this coding system. The breakage size distribution of Gaudin and Schumann equation were selected for the shredding operation simulation. The Gaudin's liberation model was suitable fur the liberation simulation. Both of these equations were transformed weight percentage into particles distribution for the necessary of particle coding method. A better recycling operation for this shredded solid waste can be concluded from the comparison of simulation results with their sorted grade, recovery or economic of materials in different processes.

• 방송공학회논문지
• /
• 제22권6호
• /
• pp.734-743
• /
• 2017

본 논문에서는 동영상의 압축 정보를 이용하여 동영상 조작 시 발생하는 특징 패턴을 분석하여 동영상의 삭제 여부를 검출하는 기법에 대해 소개한다. 제안 방식에서는 최근 표준 코덱으로 개발되어 향후 널리 사용될 것으로 예상되는 HEVC 코덱을 이용한다. 우선 조작된 동영상과 그렇지 않은 동영상의 HEVC 부호화 패턴 중 분류하기가 용이한 여러 패턴들을 분석하여 특징벡터로 선정하고, 선정된 특징벡터를 기계학습을 통해 학습하여 두 그룹 간의 분류 기준을 모델링하여 동영상에 대한 삭제 여부를 판단한다. 실험 결과, 제안한 방식이 이전의 연구 결과에 비해 HEVC 코덱 환경에서 더욱 효과적으로 삭제 여부를 판단함을 확인하였다.

• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1998년도 추계학술대회
• /
• pp.255-256
• /
• 1998

In this paper, we propose the multi-axial classification system using parallel coding method that is systemic and flexible properties for representing dental clinical behavior. The methodology and organization of this thesis as follows. First, an analysis of other classification systems. Second, the domain of medical behavior and axises using selected elements was were determined. Third, the new code system is constructed of these common factors in properties of prediction of hierarchy, brevity, simplicity, flexibility and mnemonic usage. Finally, the framework of classification system for dental was made using multi-axial code system. The result of the this study, the eight bases axis of multi-axial code system is composed and can be basic information of research for construction of classification system of all medical domain.

• 한국정밀공학회지
• /
• 제14권1호
• /
• pp.58-68
• /
• 1997

In order to provide economic high-quality products to customers in a timely manner, companies have tried much effort to decrease the time period of engineering design and information management. As a part of this effort, we have developed the Drawing Information Management System(DIMS) based ofn GT(Group Technology) that could decrease design processing time by speedy and rational management of design processes. The characteristics of DIMS are as follows: First, the concept of Concurrent Engineering was applied to DIMS. Through LAN, reviewers are able to attach comments to dlectronic documents by anno- tation functions called Mark-up. The reviewer annotations are collected and combind with the original document to revise the documents. Second, we have developed a Classification and Coding(C&C) system suitable for electronic component parts bassed on GT(Group Technology). The C&C system makes both parts and drawing with similar characteriscs into families and helps users search existing documents or create new drawings promptly. Finally, DIMS provides the Engineering BOM(Bill of Material) using the concept of Family BOM based on model options.