• 한국멀티미디어학회논문지
• /
• 제19권1호
• /
• pp.1-9
• /
• 2016

In this paper, we propose a method of measuring bone mineral density in a peripheral-type clinical X-ray CT using a phantom, and we propose a method of classifying osteoporosis using bone mineral density and bone structure parameters together. It segments the trabecular bone region and cortical bone region for the six sections of the phantom and calculates the average HU value of the segmented regions. By using these values, it derives an expression converting HU value to bone mineral density. It segments trabecular bone of 1 cm region in the end part of distal radius and extracts the bone mineral density and structural parameters for the trabecular bone region. We extracted bone mineral density and structural parameters for the 18 subjects each of normal and osteoporotic group. We carried out classification experiments using three classification methods; SAD, SVM, ANN. The sensitivity, specificity, accuracy, positive predictive value, negative predictive value, likelihood ratio of the classification was improved in the order of ANN, SVM and SAD. Also, The sensitivity, specificity, accuracy, positive predictive value, negative predictive value, likelihood ratio of the classification was improved when we use the bone mineral density and structural parameters together.

• 대한원격탐사학회지
• /
• 제28권6호
• /
• pp.611-622
• /
• 2012

This study proposed two multisensor fusion methods for segment-based image classification utilizing a region-growing segmentation. The proposed algorithms employ a Gaussian-PDF measure and an evidential measure respectively. In remote sensing application, segment-based approaches are used to extract more explicit information on spatial structure compared to pixel-based methods. Data from a single sensor may be insufficient to provide accurate description of a ground scene in image classification. Due to the redundant and complementary nature of multisensor data, a combination of information from multiple sensors can make reduce classification error rate. The Gaussian-PDF method defines a regional measure as the PDF average of pixels belonging to the region, and assigns a region into a class associated with the maximum of regional measure. The evidential fusion method uses two measures of plausibility and belief, which are derived from a mass function of the Beta distribution for the basic probability assignment of every hypothesis about region classes. The proposed methods were applied to the SPOT XS and ENVISAT data, which were acquired over Iksan area of of Korean peninsula. The experiment results showed that the segment-based method of evidential measure is greatly effective on improving the classification via multisensor fusion.

• Journal of information and communication convergence engineering
• /
• 제19권2호
• /
• pp.79-83
• /
• 2021

This paper proposes a novel image classification method based on few-shot learning, which is mainly used to solve model overfitting and non-convergence in image classification tasks of small datasets and improve the accuracy of classification. This method uses model structure optimization to extend the basic convolutional neural network (CNN) model and extracts more image features by adding convolutional layers, thereby improving the classification accuracy. We incorporated certain measures to improve the performance of the model. First, we used general methods such as setting a lower learning rate and shuffling to promote the rapid convergence of the model. Second, we used the data expansion technology to preprocess small datasets to increase the number of training data sets and suppress over-fitting. We applied the model to 10 monkey species and achieved outstanding performances. Experiments indicated that our proposed method achieved an accuracy of 87.92%, which is 26.1% higher than that of the traditional CNN method and 1.1% higher than that of the deep convolutional neural network ResNet50.

• 한국환경복원기술학회지
• /
• 제11권2호
• /
• pp.55-65
• /
• 2008

This study was surveyed to analyze vegetation structure of mountainous wetlands in Chung-ju city and around area from September to November, 2006. 6 sites of total 15 potential mountainous wetlands were discriminated throughout field survey. By classification system of mountainous wetlands presented in manual of forest wetlands research, types of wetlands were classified into slant and a flat. Many sites were covered with land plants as Pueraria thunbergiana and so on. To understand vegetation structure of mountainous wetlands, Height, DBH (diameter at breast height), DI (Dominance Index), Sociability and Constancy were surveyed and Based on this result, a projection chart was drawn. As results, Salix koreensis in tree layer and Persicaria thunbergii and Impatiens textori in herb layer were surveyed as broadly distributed species. This study is mainly focused on vegetation condition of mountainous wetlands. But, it will be needed studying on classification system of mountainous wetland type and functional assessment for conservation or management of wetlands.

• 한국환경생태학회지
• /
• 제8권2호
• /
• pp.107-120
• /
• 1995

본 연구는 주왕산국립공원내 주왕계곡의 식물군집구조를 밝히기 위하여 30개소에 조사구(1개 조사구당 100$m^2$)를 설치하여 식생조사를 실시하였으며 군집구조분석에는 상대우점치, 흉고직경급 분포, 종다양도지수, DCA, CCA 및 TWINSPAN분석을 적용하였다. 30개 조사구의 우점수종은 소나무, 졸참나무, 신갈나무, 굴참나무이었으며 전체 조사구는 TWINSPAN에 의해 신갈나무-소나무-서어나무군집, 소나무-굴참나무군집, 졸참나무-굴참나무군집, 활엽수혼효림의 4개군집으로 분리되었다. 본 조사지의 종다양도는 1.17~l.32로 높게 나타났다. 천이계열은 교목상층에서 소나무$\longrightarrow$참나무류로의 천이가 진행될 것으로 예측되었다.

• 환경정책연구
• /
• 제3권2호
• /
• pp.113-137
• /
• 2004

The objective of this study is to introduce the definitions and classification methods of wastes in international agreements and legislations, examine the concept of wastes and their classification systems in Korea, and finally analyze and compare the concept of wastes in different countries for finding better solutions and suggestions. The study summarizes the concept of wastes as introduced in the Basel Convention, OECD, EU, US, and UK. First, each of the member countries adapt to the same concepts of wastes as defined in their international agreements; second, the intention of the wastes holder and the conditions of the wastes are considered at the same time when defining the concepts. Upon close examination of the classification of wastes systems as introduced in the Basel Convention, OECD, EU, US, and UK, the wastes are classified into toxic and non-toxic wastes according to the existence of poisonous substances. Therefore, it is classified as a toxic waste when any toxic substance on its list is included in the waste, while others are considered as a non-toxic waste if they don't contain poisonous substances. Secondly, in the UK, the matter of toxic or non-toxic wastes are classified, not according to the existence of the poisonous substances, but based on the generation of sources. In Korea, the concepts of wastes are divided into the two categories - a concept as defined in actual legislations and a concept in its translation. The Korean classification of the wastes include Wastes Management Act, amended in 1995, which stipulates that toxic substances should be managed in a special way as the designated wastes. It appears that the Act utilizes the classification method that classifies the wastes according to the existence of poisonous substance. Korea's concepts of wastes should be changed after recognition of the concepts in international agreement (Basel Convention, EU) and other foreign laws(US, UK) that consider subjective and objective standards at the same time when they define the concepts. Also, the development of technology in recycling and reuse of the wastes can remove the current absolute notion of the wastes so that it also should not be passed over. Also, because a classification structure of wastes has a close relationship with a disposal structure, its classification system should be fixed gradually to come up with the development of wastes disposal technology and its policy.

• 정보관리학회지
• /
• 제18권3호
• /
• pp.285-309
• /
• 2001

인터넷의 확산에 따라 산업정보의 인터넷 자원이 급속도로 증가하고 있다. 그러나 정보서비스 기관마다 나름대로의 분류기준으로 정보를 분류하고 있는 것이 현실이다. 이로 인해 산업정보의 체계적이고 지속적인 구축은 물론 산업활동 주체간, 업무특성별 상호간에 산업정보의 활용에도 많은 불편을 초래하고 있다. 본 연구는 산업정보를 효과적으로 검색할 수 있도록 하기 위해 합리적이고 체계적인 분류체계를 제시하여 인터넷을 통한 정보접근에 익숙하지 않은 이용자도 쉽게 이용할 수 있도록 하는데 목적이 있다. 따라서 본 연구는 산업정보의 근간이라고 할 수 있는 {한국표준산업분류표}를 중심으로 현재 서비스하고 있는 국내 산업정보서비스 사이트의 주제분류체계를 조사하여 분석하고, 현재 인터넷으로 서비스하고 있는 산업정보의 양을 계량적으로 측정하여 분류항목 선정의 타당성을 분석하고, 이를 토대로 분류원칙과 분류지가 타당한 산업분류체계의 구성방안을 제시하였다.

• 전자공학회논문지CI
• /
• 제48권6호
• /
• pp.35-46
• /
• 2011

본 논문에서는 피셔 선형 분리(FLD, Fisher's Linear Discriminant) 기반의 단계적 분류를 이용한 감성 인식 기법을 제안한다. 제안하는 기법은 2종 이상의 감성에 대한 다중 클래스 분류 문제에 대하여, 이진 분류 모델의 연속적인 결합을 통해 단계적 분류 모델을 구성함으로써 복잡도 높은 특징 공간상의 다수의 감성 클래스에 대한 분류 성능을 향상시킨다. 이를 위하여, 각 계층 단계의 학습에서는 감성 클래스들로 이루어진 두 개의 클래스 그룹에 따라 피셔 선형분리 공간을 구성하며, 구성된 공간상에서 Adaboost 방식을 이용하여 이진 분류 모델을 학습하여 생성한다. 각 계층 단계의 학습 과정은 모든 감성 클래스가 구분이 완료되는 시점까지 반복 수행된다. 본 논문에서는 MIT 생체 신호 프로파일을 이용하여 제안하는 기법을 실험하였다. 실험 결과, 8종의 감성에 대한 분류 실험을 통해 약 72%의 분류 성능을 확인하였고, 특정 3종의 감성에 대한 분류 실험을 통해 약 93% 분류 성능을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권7호
• /
• pp.3172-3193
• /
• 2018

Recently, several studies have shown that linear representation based approaches are very effective and efficient for image classification. One of these linear-representation-based approaches is the Collaborative representation (CR) method. The existing algorithms based on CR have two major problems that degrade their classification performance. First problem arises due to the limited number of available training samples. The large variations, caused by illumintion and expression changes, among query and training samples leads to poor classification performance. Second problem occurs when an image is partially noised (contiguous occlusion), as some part of the given image become corrupt the classification performance also degrades. We aim to extend the collaborative representation framework under limited training samples face recognition problem. Our proposed solution will generate virtual samples and intra-class variations from training data to model the variations effectively between query and training samples. For robust classification, the image patches have been utilized to compute representation to address partial occlusion as it leads to more accurate classification results. The proposed method computes representation based on local regions in the images as opposed to CR, which computes representation based on global solution involving entire images. Furthermore, the proposed solution also integrates the locality structure into CR, using Euclidian distance between the query and training samples. Intuitively, if the query sample can be represented by selecting its nearest neighbours, lie on a same linear subspace then the resulting representation will be more discriminate and accurately classify the query sample. Hence our proposed framework model the limited sample face recognition problem into sufficient training samples problem using virtual samples and intra-class variations, generated from training samples that will result in improved classification accuracy as evident from experimental results. Moreover, it compute representation based on local image patches for robust classification and is expected to greatly increase the classification performance for face recognition task.

• 정보교육학회논문지
• /
• 제21권3호
• /
• pp.297-307
• /
• 2017

본 연구의 목적은 초중등학생 대상 알고리즘 교육을 위한 알고리즘 분류체계를 제안하는 것이다. 연구자는 알고리즘의 구성요소를 정의하고, 분석합성식 방법으로 알고리즘 분류체계를 표현하였다. 연구의 내용은 다음과 같다. 첫째, 분류의 목적과 분류의 종류에 대한 이론적인 탐색을 실시하였다. 둘째, 기존에 제안된 알고리즘 내용에 대한 분류체계의 내용과 그 한계에 대해 살펴보았다. 이와 더불어 알고리즘 교육 연구에서 사용되었던 알고리즘 교육 내용 및 선정 기준에 대해 살펴보았다. 셋째, 알고리즘의 분류를 위해 알고리즘 구성요소를 NRC에서 제시한 핵심 아이디어와 관통 개념을 사용하여 재정의하였다. 그리고 알고리즘 관통 개념을 디자인 구조와 자료구조로 세분화하여 그 내용을 제시하였으며, 이 내용을 분석합성식 분류체계를 사용하여 표현하였다. 마지막으로 전문가 집단의 검토를 통해 제안한 내용에 대한 타당도를 검증하였다. 알고리즘 분류체계에 대한 연구는 알고리즘 교육에 있어 내용 선정 및 교육 방법에 많은 시사점을 제공할 것으로 기대한다.