• Proceedings of the IEEK Conference
• /
• 2002.06a
• /
• pp.251-254
• /
• 2002

본 연구에서는 G.983.1 표준에 WDM (Wavelength Division Multiolexing) 기술을 이용하여 WDM-PON 구조를 제시하고, PON상에서의 주된 지연은 upstream 상에서 발생하므로 IEEE 802.Ip를 토대로 각각의 클래스별로 패킷을 분류하여 제안한 알고리즘을 통해 상향 링크상에서의 동적 채널 할당을 실험 하였다. 이를 토대로 WDM-PON에서 각 클래스별 다중 버퍼 방식의 효율성과 채널 할당 알고리즘의 타당성을 확인해 보고자 한다.

• Journal of Korea Multimedia Society
• /
• v.11 no.7
• /
• pp.1035-1042
• /
• 2008

The support vector machine, as a binary classifier, is known to surpass the other classifiers only in binary classification problems through the various experiments. Even though its theory is based on the maximal margin classifier, the support vector machine approach cannot be easily extended to the multi-classification problems. In this paper, we review the extension techniques of the support vector machine toward the multi-classification and do the performance comparison. Depending on the data decomposition of the training data, the support vector machine is easily adapted for a multi-classification problem without modifying the intrinsic characteristics of the binary classifier. The performance is evaluated on a collection of the benchmark data sets and compared according to the selected teaming strategies, the training time, and the results of the neural network with the backpropagation teaming. The experiments suggest that the support vector machine is applicable and effective in the general multi-class classification problems when compared to the results of the neural network.

• Journal of the Korea Convergence Society
• /
• v.10 no.9
• /
• pp.23-29
• /
• 2019

In this paper, the mood of images was classified into eight categories through a deep convolutional neural network and video was automatically generated using proper background music. Based on the collected image data, the classification model is learned using a multilayer perceptron (MLP). Using the MLP, a video is generated by using multi-class classification to predict image mood to be used for video generation, and by matching pre-classified music. As a result of 10-fold cross-validation and result of experiments on actual images, each 72.4% of accuracy and 64% of confusion matrix accuracy was achieved. In the case of misclassification, by classifying video into a similar mood, it was confirmed that the music from the video had no great mismatch with images.

• The KIPS Transactions:PartB
• /
• v.16B no.4
• /
• pp.281-288
• /
• 2009

This paper proposes a novel algorithm for the efficient classification and annotation of medical images, especially X-ray images. Since X-ray images have a bright foreground against a dark background, we need to extract the different visual descriptors compare with general nature images. In this paper, a Color Structure Descriptor (CSD) based on Harris Corner Detector is only extracted from salient points, and an Edge Histogram Descriptor (EHD) used for a textual feature of image. These two feature vectors are then applied to a multi-class Support Vector Machine (SVM), respectively, to classify images into one of 20 categories. Finally, an image has the Annotation Code Array based on the pre-defined hierarchical relations of categories and priority code order, which is given the several optimal keywords by the Annotation Code Array. Our experiments show that our annotation results have better annotation performance when compared to other method.

• Journal of Korean Society for Geospatial Information Science
• /
• v.24 no.1
• /
• pp.121-128
• /
• 2016

Many studies are focused on image data and classifier for comparison or improvement of classification accuracy. Therefore studies are needed aspect of the training samples on supervised classification which depend on reference data or skill of analyst. This study tries to assess usability of field spectra as training samples on supervised classification. Classification accuracies of hyperspectral and multispectral images were assessed using training samples from image itself and field spectra, respectively. The results shown about 90% accuracy with training sample collected from image. Using field spectra as training sample, accuracy was decreased 10%p for hyperspectral image, and 20%p for multispectral image. Especially, some classes shown very low accuracies due to similar spectral characteristics on multispectral image. Therefore, field spectra might be used as training samples on classification of hyperspectral image, although it has limitation for multispectral image.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.1
• /
• pp.27-35
• /
• 2021

In this paper, we propose a new stacking ensemble framework for deep learning models which reflects the distribution of label embeddings. Our ensemble framework consists of two phases: training the baseline deep learning classifier, and training the sub-classifiers based on the clustering results of label embeddings. Our framework aims to divide a multi-class classification problem into small sub-problems based on the clustering results. The clustering is conducted on the label embeddings obtained from the weight of the last layer of the baseline classifier. After clustering, sub-classifiers are constructed to classify the sub-classes in each cluster. From the experimental results, we found that the label embeddings well reflect the relationships between classification labels, and our ensemble framework can improve the classification performance on a CIFAR 100 dataset.

• Proceedings of the KSRS Conference
• /
• 2006.03a
• /
• pp.210-213
• /
• 2006

이 연구에서는 다중시기 Landsat 영상을 이용하여 강원도 일대 임상의 변화를 살펴보고 상록수의 영급을 구분하는 알고리즘을 개발하여 적용하였다. 1980년대에서 현재까지 축적된 Landsat-5와 Landsat-7영상 중에서, 대부분 지역에 활잡목 및 활엽수가 낙엽이 지고 눈이 아직 쌓이지 않을 시기인 11월에 촬영된 영상만을 이용하였다. 각 영상에서 양지바른 상록수, 활엽수, 그늘진 지역, 도시 및 바다 등을 클래스로 지정하여 감돌분류를 하였다. 분류 결과에서 양지바른 상록수만 추출하여 5개의 영상을 이진 분류체계로 조합한 후 임상의 시기적 변화 양상을 관찰한 결과, 강원대 연습림의 조림 기록 및 현황도와 상당히 일치함을 확인하였으며, Path 115, Row 34에 해당하는 강원도 일대로 연구지역을 확대하였다. 향후 Kompsat-2를 비롯한 고해상도 11월 영상이 지속적으로 촬영된다면, 이 연구에서 개발된 이진 분류체계 방법을 통하여 산림변화의 모니터링을 보다 용이하고 효율적으로 할 수 있을 것으로 기대된다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2007.10a
• /
• pp.38-39
• /
• 2007

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.37 no.2
• /
• pp.39-49
• /
• 2000

In this paper, we propose an optimizing feature extraction method for multiclass problems assuming normal distributions. Initially, We start with an arbitrary feature vector Assuming that the feature vector is used for classification, we compute the classification error Then we move the feature vector slightly in the direction so that classification error decreases most rapidly This can be done by taking gradient We propose two search methods, sequential search and global search In the sequential search, an additional feature vector is selected so that it provides the best accuracy along with the already chosen feature vectors In the global search, we are not constrained to use the chosen feature vectors Experimental results show that the proposed algorithm provides a favorable performance.

• Journal of the Society of Disaster Information
• /
• v.14 no.4
• /
• pp.501-509
• /
• 2018

Purpose: UAV-based photo measurements are being researched using UAVs in the space information field as they are not only cost-effective compared to conventional aerial imaging but also easy to obtain high-resolution data on desired time and location. In this study, the UAV-based high-resolution images were used to perform the land cover classification. Method: RGB cameras were used to obtain high-resolution images, and in addition, multi-distribution cameras were used to photograph the same regions in order to accurately classify the feeding areas. Finally, Land cover classification was carried out for a total of seven classes using created ortho image by RGB and multispectral camera, DSM(Digital Surface Model), NDVI(Normalized Difference Vegetation Index), GLCM(Gray-Level Co-occurrence Matrix) using RF (Random Forest), a representative supervisory classification system. Results: To assess the accuracy of the classification, an accuracy assessment based on the error matrix was conducted, and the accuracy assessment results were verified that the proposed method could effectively classify classes in the region by comparing with the supervisory results using RGB images only. Conclusion: In case of adding orthoimage, multispectral image, NDVI and GLCM proposed in this study, accuracy was higher than that of conventional orthoimage. Future research will attempt to improve classification accuracy through the development of additional input data.