• Journal of Korean Association for Spatial Structures
• /
• v.14 no.4
• /
• pp.55-64
• /
• 2014

In this study, Intersecting Tensegrity System that is integrated solid compression members with tension members was presented. This system is set up by connecting upper and lower compression members of pyramid shape with exterior tension members. In this system, the solid compression members are intersected each other and connected by a tension member in the center. This system is a variation of Tensegrity system, has a improved feature that the system is able to induce prestresses in all of tension members easily by adjusting the distance of a tension member in the center. The proposed system was studied by modeling, and the structural behavior of the system was investigated by mechanical analysis of the model. Furthermore, the features of the structural behavior variations was investigated when the composition elements(total height, size of surface, intersection length, etc.) are changed variously. It was also showed that the system is able to be used as a temporary space structure system with a membrane roof of inverse conical shape.

• Journal of Korean Association for Spatial Structures
• /
• v.13 no.1
• /
• pp.87-96
• /
• 2013

This study investigated characteristics of buckling load and effective buckling length by member rigidity of dome-typed space frame which was sensitive to initial conditions. A critical point and a buckling load were computed by analyzing the eigenvalues and determinants of the tangential stiffness matrix. The hexagonal pyramid model and star dome were selected for the case study in order to examine the nodal buckling and member buckling in accordance with member rigidity. From the numerical results, an effective buckling length factor of adopted models was bigger than that of Euler buckling for the case of fixed boundary. These numerical models indicated that the influence of nodal buckling was greater than that of member buckling as member rigidity was higher. Besides, there was a tendency that the bifurcation appeared on the equilibrium path before limit point in the member buckling model.

• Korean Journal of Remote Sensing
• /
• v.17 no.1
• /
• pp.57-69
• /
• 2001

This study propose a image processing system of unsupervised analysis. This system integrates low-level segmentation and high-level classification. The segmentation and classification are conducted respectively with and without spatial constraints on merging by a hierarchical clustering procedure. The clustering utilizes the local mutually closest neighbors and multi-window operation of a pyramid-like structure. The proposed system has been evaluated using simulated images and applied for the LANDSATETM+ image collected from Youngin-Nungpyung area on the Korean Peninsula.

• Journal of Korea Multimedia Society
• /
• v.22 no.3
• /
• pp.335-348
• /
• 2019

Cell segmentation and counting represent one of the most important tasks required in order to provide an exhaustive understanding of biological images. Conventional features suffer the lack of spatial consistency by causing the joining of the cells and, thus, complicating the cell counting task. We propose, in this work, a cascade of networks that take as inputs different versions of the original image. After constructing a Gaussian pyramid representation of the microscopy data, the inputs of different size and spatial resolution are given to a cascade of deep convolutional autoencoders whose task is to reconstruct the segmentation mask. The coarse masks obtained from the different networks are summed up in order to provide the final mask. The principal and main contribution of this work is to propose a novel method for the cell counting. Unlike the majority of the methods that use the obtained segmentation mask as the prior information for counting, we propose to utilize the hidden latent representations, often called the high-level features, as the inputs of a neural network based regressor. While the segmentation part of our method performs as good as the conventional deep learning methods, the proposed cell counting approach outperforms the state-of-the-art methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.3
• /
• pp.1614-1632
• /
• 2017

Bag of visual words is a popular model in human action recognition, but usually suffers from loss of spatial and temporal configuration information of local features, and large quantization error in its feature coding procedure. In this paper, to overcome the two deficiencies, we combine sparse coding with spatio-temporal pyramid for human action recognition, and regard this method as the baseline. More importantly, which is also the focus of this paper, we find that there is a hierarchical structure in feature vector constructed by the baseline method. To exploit the hierarchical structure information for better recognition accuracy, we propose a tree regularized classifier to convey the hierarchical structure information. The main contributions of this paper can be summarized as: first, we introduce a tree regularized classifier to encode the hierarchical structure information in feature vector for human action recognition. Second, we present an optimization algorithm to learn the parameters of the proposed classifier. Third, the performance of the proposed classifier is evaluated on YouTube, Hollywood2, and UCF50 datasets, the experimental results show that the proposed tree regularized classifier obtains better performance than SVM and other popular classifiers, and achieves promising results on the three datasets.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2011.02a
• /
• pp.47-47
• /
• 2011

There have been numerous efforts to enhance the efficiency of light-emitting diodes (LEDs) by using low dimensional structures such as quantum dots (QDs), wire (QWRs), and wells (QWs). We demonstrate QD/QWR/QW hybrid structured LEDs by using nano-scaled pyramid structures of GaN with ~260 nm height. Photoluminescence (PL) showed three multi-peak spectra centered at around 535 nm, 600 nm, 665 nm for QWs, QWRs, and QDs, respectively. The QD emission survived at room temperature due to carrier localization, whereas the QW emission diminished from 10 K to 300 K. We confirmed that hybrid LEDs had zero-, one-, and two-dimensional behavior from a temperature-dependent time-resolved PL study. The radiative lifetime of the QDs was nearly constant over the temperature, while that of the QWs increased with increasing temperature, due to low dimensional behavior. Cathodoluminescence revealed spatial distributions of InGaN QDs, QWRs, and QWs on the vertices, edges, and sidewalls, respectively. We investigated the blue-shifted electroluminescence with increasing current due to the band-filling effect. The hybrid LEDs provided broad-band spectra with high internal quantum efficiency, and color-tunability for visible light-emitting sources.

• Journal of Korean Society for Geospatial Information Science
• /
• v.16 no.2
• /
• pp.49-56
• /
• 2008

Image registration is one of the critical techniques of image mosaic which has many applications such as generating panoramas, video monitoring, image rendering and reconstruction, etc. The fundamental tasks of image registration are point features extraction and tracking which take much computation time. KLT(Kanade-Lucas-Tomasi) feature tracker has proposed for extracting and tracking features through image sequences. The aim of this study is to demonstrate the usage of effective and robust KLT feature detector and tracker for an image registration using the sequence image frames captured by UAV video camera. In result, by using iterative implementation of the KLT tracker, the features extracted from the first frame of image sequences could be successfully tracked through all frames. The process of feature tracking in the various frames with rotation, translation and small scaling could be improved by a careful choice of the process condition and KLT pyramid implementation.

• Korean Journal of Optics and Photonics
• /
• v.16 no.2
• /
• pp.113-120
• /
• 2005

Luminous efficiency and uniformity in a LCD-BLU are mainly determined by fine scattering patterns formed on the light guide panel. We propose a novel fabrication method of 3-dimensional scattered patterns based on anisotropic etching of silicon wafers. Micro-pyramid patterns with 70.5 degree apex-angle and micro-prism patterns with 109.4 degree apex-angle can be self-constructed by the wet, anisotropic etching of (100) and (110) silicon wafers, respectively, and those patterns are easily duplicated by the PDMS replica process. Experimental results on spatial and angular distributions of irradiation from the light guide panel with the micro-pyramid patterns were very consistent with the calculation results. Surface roughness of the silicon-based micro-patterns is free from any artificial defects since the micro-patterns are inherently formed with silicon crystal surfaces. Therefore, we expect that the silicon based micro-patterning process makes it possible to fabricate perfect 3-dimensional micro-structures with crystal surface and apex angles, which may guarantee mass-reproduction of the light guide panels in LCD-BLU.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1057-1060
• /
• 2004

We propose new rotation moment invariants based on multiresolution filter bank techniques. The multiresolution pyramid motivates our simple but efficient feature selection procedure based on the fuzzy C-mean clustering, combined with the Mahalanobis distance. The procedure verifies an impact of random noise as well as an interesting and less known impact of noise due to spatial transformations. The recognition accuracy of the proposed techniques has been tested with the preceding moment invariants as well as with some wavelet based schemes. The numerical experiments, with more than 30,000 images, demonstrate a tangible accuracy increase of about 3% for low noise, 8% for the average noise and 15% for high level noise.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.3
• /
• pp.861-880
• /
• 2023

Semantic segmentation of road scene is the key technology of autonomous driving, and the improvement of convolutional neural network architecture promotes the improvement of model segmentation performance. The existing convolutional neural network has the simplification of learning knowledge and the complexity of the model. To address this issue, we proposed a road scene semantic segmentation algorithm based on multi-task collaborative learning. Firstly, a depthwise separable convolution atrous spatial pyramid pooling is proposed to reduce model complexity. Secondly, a collaborative learning framework is proposed involved with saliency detection, and the joint loss function is defined using homoscedastic uncertainty to meet the new learning model. Experiments are conducted on the road and nature scenes datasets. The proposed method achieves 70.94% and 64.90% mIoU on Cityscapes and PASCAL VOC 2012 datasets, respectively. Qualitatively, Compared to methods with excellent performance, the method proposed in this paper has significant advantages in the segmentation of fine targets and boundaries.