• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2008.11a
• /
• pp.255-257
• /
• 2008

본 논문에서는 최근 IT 기술의 발전에 따라 무수히 양산되고 있는 멀티미디어 데이터를 효율적으로 검색하기 위한 방법을 제안한다. 영상 검색 시스템에 사용되는 데이터베이스(DB) 영상들에 존재하는 각 객체들의 존재 영역을 기반으로 질의 영상 (query image)의 객체 영역을 추정해서 검색에 활용하는 것이다. 이는 질의 영상의 전체 영역으로부터 객체를 추정하는 것보다 데이터베이스 영상들로부터 추출한 통계적 객체 분포 범위를 기반으로 추정하기 때문에 빨리 객체 추출이 가능하도록 한다. 따라서 객체를 추출하기 위한 배경 지식이나, 사용자 입력이 전혀 필요 없다. 이렇게 추출된 객체 영역의 영상들로부터 GLCM 알고리즘을 이용해서 객체 영역의 특성이 잘 반영된 질감 특징 값을 바탕으로 검색에 활용 할 경우 원본 영상의 질감 특징을 활용한 경우보다, 객체의 질감 특징을 더 잘 반영한다는 것을 실험을 통해 확인할 수 있었다.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2000.04a
• /
• pp.328-331
• /
• 2000

In this paper , we propose a texture feature extraction method of reduce the massive computational time on extracting texture, features of large sized medical such as MRI, CT-scan , and an index structure, called GLTFT, to speed up the retrieval performance. For these, the original image is transformed into a compressed image by Wavelet transform , and textural features such as contrast, energy, entropy, and homogeneity of the compressed image is extracted by using GLCM(Gray Level Co-occurrence Metrix) . The proposed index structure is organized by using the textural features. The processing in compressed domain can give the solution of storage space and the reduction of computational time of feature extracting . And , by GLTFT index structure, image retrieval performance can be expected to be improved by reducing the retrieval range . Our experiment on 270 MRIs as image database shows that shows that such expectation can be got.

• Proceedings of the KOSOMBE Conference
• /
• v.1990 no.11
• /
• pp.25-28
• /
• 1990

We tested the ability of two-dimensional echocardiograms of complete heart cycle in closed-chest human to discriminate between normal and infarcted myocardium using fixed window, Inter- and Intra-frame analysis. The results show that statistical parameter, MEAN, second order gray level statistics parameter, ASM and proposed parameter, HGE, I.T, can quantitatively distinguish between normal and Infarcted regions. The manner in which these parameters vary over the cardiac cycle is also a good indicator of the state of myocardium. The infarcted areas yield regions of higher Intensity throughout the cardiac cycle. Whereas, normal tissue demonstrates greater variability throughout the cardiac cycle.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.4
• /
• pp.2109-2123
• /
• 2017

This paper presents a super-resolution reconstruction method based on sparse representation for two-dimensional bar code images. Considering the features of two-dimensional bar code images, Kirsch and LBP (local binary pattern) operators are used to extract the edge gradient and texture features. Feature extraction is constituted based on these two features and additional two second-order derivatives. By joint dictionary learning of the low-resolution and high-resolution image patch pairs, the sparse representation of corresponding patches is the same. In addition, the global constraint is exerted on the initial estimation of high-resolution image which makes the reconstructed result closer to the real one. The experimental results demonstrate the effectiveness of the proposed algorithm for two-dimensional bar code images by comparing with other reconstruction algorithms.

• Proceedings of the KSRS Conference
• /
• 2007.10a
• /
• pp.247-249
• /
• 2007

Crown parameters are important roles in tree species identification, because the canopy is the aggregate of all the crowns. However, crown measurements with spaceborne image data have remained more difficult than on aerial photographs since trees show more structural detail at higher resolutions. This recognized problem led to the initiation of the research to determine if high resolution satellite image data could be used to identify and classify single tree species. In this paper, shape parameters derived from pixel-based crown area measurements and texture features derived from GLCM parameters in QuickBird image were tested and compared for individual tree species identification. As expected, initial studies have shown that the crown parameters and the canopy texture parameters provided a differentiating method between coniferous trees and broad-leaved trees within the compartment(less than forest stand) for single extraction from spaceborne high resolution image.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.12
• /
• pp.51-58
• /
• 1994

Mathematical morphology based on set theory is easy to be implemented in parallel and can be applied to various fields in image analysis. Particularly mophological pattern spectrum can detect critical scales in an image object and quantify various aspects of the shape-size content. In this paper, texture classification using pattern spectrum based on morphological subband decomposition is porposed. The low-low band extracts pattern spectrum features, and the high-low, low-high, and high-high bands extrack the structural information. This approach has the advantages of efficient information extraction, less time-consuming, high accuacy, less computation, and parallel implementation.

• Korean Journal of Optics and Photonics
• /
• v.16 no.3
• /
• pp.275-286
• /
• 2005

The performance of the InGaN LED's in terms of the photon extraction efficiency has been analyzed by the Monte Carlo photon simulation method. Simulation results show that the sidewall slanting scheme, which works well for the AlInGaP or InGaN/SiC LED, plays a very minimal role in InGaN/sapphire LED's. In contrast to InGaN/SiC LED's, the lower refractive index sapphire substrate restricts the generated photons to enter the substrate, minimizing the chances for the photons to be deflected by the slanted sidewalls of the epitaxial semiconductor layers that are usually very thin. The limited photon transmission to the sapphire substrate also degrades the. photon extraction efficiency especially in the epitaxial-side down mount. One approach to exploit the photon extraction potential of the epitaxial-side down mount may be to texture the substrate-epitaxy interface. In this case, randomized photon deflection off the textured interface directly increases the number of the photons entering the sapphire substrate, from which they easily couple out of the chip and thereby improving the photon extraction efficiency drastically.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.245-245
• /
• 2005

This paper addresses the use of directional textural information for connecting road segments. In urban scene, some roads are occluded by buildings, casting shadow of buildings, trees, and cars on streets. Automatic extraction of road network from remotely sensed high resolution imagery is generally hindered by them. The results of automatic road network extraction will be incomplete. To overcome this problem, several perceptual grouping algorithms are often used based on similarity, proximity, continuation, and symmetry. Roads have directions and are connected to adjacent roads with certain angles. The directional information is used to guide road fragments connection based on roads directional inertia or characteristics of road junctions. In the primitive stage, roads are extracted with textural and direction information automatically with certain length of linearity. The primitive road fragments are connected based on the directional information to improve the road network. Experimental results show some contribution of this approach for completing road network, specifically in urban area.

• Journal of the Korean Society of Industry Convergence
• /
• v.23 no.3
• /
• pp.515-521
• /
• 2020

Deep learning methods have been effectively used to provide great improvement in various research fields such as machine learning, image processing and computer vision. One of the most frequently used deep learning methods in image processing is the convolutional neural networks. Compared to the traditional artificial neural networks, convolutional neural networks do not use the predefined kernels, but instead they learn data specific kernels. This property makes them to be used as feature extractors as well. In this study, we compared the quality of CNN features for traditional texture feature extraction methods. Experimental results demonstrate the superiority of the CNN features. Additionally, the recognition process and result of a pioneering CNN on MNIST database are presented.

• Journal of the Korean Society of Food Science and Nutrition
• /
• v.30 no.3
• /
• pp.488-493
• /
• 2001

The know the possibility in development of the low-fat beef patty models using crucian carp9 extraction residues (CCER, freeze dried powder : 5%, 10%, 15%), those protein nutritional quality, texture, color and sensory properties were investigated. About 13∼23% (on dry basis) of lipid in control was reduced in cooked beef patties with the higher addition ratios of CCER. In vitro protein digestibility was not changed in raw patties before cooking but 2∼4% higher digestibility was revealed in cooked patties. Computed protein efficiency ratio (C-PER) and discriminant computed protein efficiency ratio (DC-PER) of beef patties containing CCER were almost same as control. Lightness and red color value of both (raw and cooked) beef patties were decreased with the higher CCER addition ratios but brown color value of cooked samples were similar to control. Stronger hardness was noted in all beef patties containing CCER significantly (p<0.05). Consumer's acceptability were generally decreased by addition of CCER, but 10% level could be recommendable in beef patty processing.