• Proceedings of the Korean Society of Computer Information Conference
• /
• 2022.07a
• /
• pp.23-26
• /
• 2022

딥 러닝 모델은 블랙 박스 (Black Box) 모델로 예측에 대한 근거를 제시하지 못해 신뢰성이 떨어지는 단점이 존재한다. 이를 해결하기 위해 딥 러닝 모델에 설명력을 부여하는 설명 가능한 인공지능 (XAI) 분야 연구가 활발하게 이루어지고 있다. 본 논문에서는 모델 예측을 프로토타입을 통해 설명하는 딥 러닝 모델을 제시한다. 즉, "주어진 이미지는 티셔츠인데, 그 이유는 티셔츠를 대표하는 모양의 프로토타입과 닮았기 때문이다."의 형태로 딥 러닝 모델을 설명한다. 해당 모델은 Encoder, Prototype Layer, Classifier로 구성되어 있다. Encoder는 Feature를 추출하는 데 활용하고 Classifier를 통해 분류 작업을 수행한다. 모델이 제시하는 분류 결과를 설명하기 위해 Prototype Layer에서 가장 유사한 프로토타입을 찾아 설명을 제시한다. 실험 결과 프로토타입 생성 기반 설명 모델은 기존 이미지 분류 모델과 유사한 예측 정확도를 보였고, 예측에 대한 설명력까지 확보하였다.

• Annual Conference of KIPS
• /
• 2007.05a
• /
• pp.949-950
• /
• 2007

프로토타이핑이란 개발자들과 사용자들의 의사소통상의 효과를 증진시키기 위하여 취하는 시스템개발상의 기법이며, 이러한 프로토타이핑에는 가상 프로토타이핑과 실물 프로토타이핑이 있다. 본 논문에서는 기존의 실물/가상 프로토타이핑 연동 기법을 개선해서 오브젝트 기반의 임베디드 시스템 통합 프로토타이핑 개발 환경을 구축하였다. 이러한 새로운 임베디드 시스템 개발 환경을 구축함으로써, 재사용성, 유연성, 확장성 그리고 편리성을 가질 수 있다.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.6 no.2
• /
• pp.787-802
• /
• 1964

• The KIPS Transactions:PartB
• /
• v.16B no.4
• /
• pp.263-270
• /
• 2009

Histogram Equalization (HE) is a very popular technique for enhancing the contrast of an image. HE stretches the dynamic range of an image using the cumulative distribution function of a given input image, therefore improving its contrast. However, HE has a well-known problem : when HE is applied for the contrast enhancement, there is a significant change in brightness. To resolve this problem, we propose An Adaptive Contrast Enhancement Algorithm using Subhistogram Area-Ratioed Histogram Redistribution, a new method that helps reduce excessive contrast enhancement. This proposed algorithm redistributes the dynamic range of an input image using its mean luminance value and the ratio of sub-histogram area. Experimental results show that by this redistribution, the significant change in brightness is reduced effectively and the output image is able to preserve the naturalness of an original image even if it has a poor histogram distribution.

• Journal of Korean Library and Information Science Society
• /
• v.47 no.1
• /
• pp.175-198
• /
• 2016

This study intends to implement a OPAC prototype for digital kids. To the end, this study applies prototyping techniques that design, evaluate and improve an early sample model. Included in this study are as follows: (1) design a conceptual model of new OPACs based on the information need analysis of digital kids, (2) implement a OPAC prototype with WordPress 4.2 version, (3) evaluate the prototype OPAC from the viewpoint of digital kids and children librarians, and finally, (4) suggest the improved model and implementational strategies for kid's OPAC. 6 fourth and fifth graders in the elementary school and 5 children librarians in public libraries participate in the final evaluation process.

• Journal of Korea Multimedia Society
• /
• v.15 no.6
• /
• pp.748-760
• /
• 2012

In case brightness distribution is concentrated in a region, it is difficult to classify the image features. To solve this problem, we apply global histogram equalization and local histogram equalization to images. In case of global histogram equalization, it can be too bright or dark because it doesn't consider the density of brightness distribution. Thus, it is difficult to enhance the local contrast in the images. In case of local histogram equalization, it can produce unexpected blocks in the images. In order to enhance the contrast in the images, this paper proposes a local histogram equalization based on the Gaussian Mixture Models(GMMs) in regions of histogram. Mean and variance parameters in each regions is updated EM-algorithm repeatedly and then ranges of equalization on each regions. The experimental results performed with image of various contrasts show that the proposed algorithm is better than the global histogram equalization.

• Archives of design research
• /
• v.13 no.2
• /
• pp.103-109
• /
• 2000

Rapid prototyping technology has been widely applied to the design process in the industry. This technology made it easy to generate a prototype which acts like a real product in logic. However, this technology has not been validated sufficiently as a usability test method. The purpose of this study was to test the hypothesis : There was no significant difference between a real product and its prototype. An experiment was designed to test it statistically. a microwave oven was selected and its prototype was generated using a rapid prototyping tool. Six subjects used the microwave oven and another six subjects used the prototype to perform five scenario tasks. As a result, there was significant difference between the real product and prototype in success rate, task completion time, and number of buttons pressed. The prototype was more difficult to operate than the real product. Therefore we should be careful when we apply rapid prototyping. technology. In the discussions, the causes of the difference were identified and some guidelines were suggested for who wants to apply rapid prototyping tool to the usability test.

• Journal of KIISE:Databases
• /
• v.30 no.2
• /
• pp.176-183
• /
• 2003

Various techniques including histograms, sampling and parametric techniques have been proposed to estimate query result sizes for the query optimization. Histogram-based techniques are the most widely used form for the selectivity estimation in relational database systems. However, in the spatio-temporal databases for the moving objects, the continual changes of the data distribution suffer the direct utilization of the state of the art histogram techniques. Specifically for the future queries, we need another methodology that considers the updated information and keeps the accuracy of the result. In this paper we propose a novel approach based upon the duality and the marginal distribution to construct a histogram with very little time since the spatio-temporal histogram requires the data distribution defined by query predicates. We use data synopsis method in the dual space to construct spatio-temporal histograms. Our method is robust to changing data distributions during a certain period of time while the objects keep the linear movements. An additional feature of our approach supports the dynamic update incrementally and maintains the accuracy of the estimated result.

• Journal of Korean Society of Forest Science
• /
• v.95 no.6
• /
• pp.631-635
• /
• 2006

Single-strand conformation polymorphism (SSCP) analysis of MD and JWB phytopalsma isolates which amplified PCR products using the R16F2n/R2 phytoplamsa universal primer pair were compared for variations of their nucleotide sequence. The MD and JWB phytoplasmas were clearly distinct each of the band patterns from about 1.2 kb PCR products. To clearly distinct of close SSCP band patterns, the MD and JWB phytoplasma PCR products were mixed and performed to detect their polymorphism. The SSCP band patterns show all of bands of MD and JWB on single lane and easily distinct their each band patterns. The PCR-SSCP analysis was possible to detect of 1.2 kb nucleotide sequence and near close band patterns were easily distinct by mixing two samples.

• Korean Journal of Soil Science and Fertilizer
• /
• v.17 no.2
• /
• pp.141-146
• /
• 1984

Variation of soil characteristics and purity were examined within mapping unit and estimates were given for the number of samples required to reach predetermined levels of accuracy for black soil, VDBr. soil, DBr. soil, and brown soil in Jeju. Variation of chemical properties except pH and CEC was higher than that of physical properties. Brown soil showed the lowest variation and hence required the smallest sample size. The soils examined were correctly classified with regard to brown soil at purity 74.1, to DBr. soil 60.8, to black soil 55.6, and to VDBr. soil 46.9%. Sample calculation was presented which showed approximate sample sizes required to detect the difference of 10 and 20% of the mean with 0.95 confidence level.