• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.26 no.6
• /
• pp.893-901
• /
• 2002

Effect of boundary layer thickness on the flow characteristics around a rectangular prism has been investigated by using a PIV(Particle Image Velocimetry) technique. Three different boundary layers(thick, medium and thin)were generated in the Atmospheric Boundary Layer Wind Tunnel at Pusan National University. The thick boundary layer having 670 mm thickness was generated by using spires and roughness elements. The medium thickness of boundary layer($\delta$=270 mm) was the natural turbulent boundary layer at the test section floor with fairly long developing length(18 m). The thin boundary layer($\delta$=36.5 mm) was generated on the smooth panel elevated 70cm from the wind tunnel floor. The Reynolds number based on the free stream velocity(3 ㎧) and the height of the model(40 mm) was 7.9$\times$10$^3$. The mean velocity vector fields and turbulent kinetic energy distributions were measured and compared. The effect of boundary layer thickness was clearly observed not only in the length of separation bubble but also in the location of reattachment point. The thinner the boundary layer thickness, the higher the turbulent kinetic energy Peak around the model roofbecame. It is strongly recommended that the height ratio between the model and the approaching boundary layer thickness should be encountered as a major parameter.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.4
• /
• pp.203-210
• /
• 2017

In this paper, we propose an effective neural network model for image caption generation and model transfer. This model is a kind of multi-modal recurrent neural network models. It consists of five distinct layers: a convolution neural network layer for extracting visual information from images, an embedding layer for converting each word into a low dimensional feature, a recurrent neural network layer for learning caption sentence structure, and a multi-modal layer for combining visual and language information. In this model, the recurrent neural network layer is constructed by LSTM units, which are well known to be effective for learning and transferring sequence patterns. Moreover, this model has a unique structure in which the output of the convolution neural network layer is linked not only to the input of the initial state of the recurrent neural network layer but also to the input of the multimodal layer, in order to make use of visual information extracted from the image at each recurrent step for generating the corresponding textual caption. Through various comparative experiments using open data sets such as Flickr8k, Flickr30k, and MSCOCO, we demonstrated the proposed multimodal recurrent neural network model has high performance in terms of caption accuracy and model transfer effect.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.4
• /
• pp.935-944
• /
• 2013

In this paper, a new illumination compensation algorithm by segmentation with depth information is proposed to improve the coding efficiency of multi-view images. In the proposed algorithm, a reference image is first segmented into several layers where each layer is composed of objects with a similar depth value. Then we separate objects from each other even in the same layer by labeling each separate region in the layered image. Then, the labeled reference depth image is converted to the position of the distortion image view by using 3D warping algorithm. Finally, we apply an illumination compensation algorithm to each of matched regions in the converted reference view and distorted view. The occlusion regions that occur by 3D warping are also compensated by a global compensation method. Through experimental results, we are able to confirm that the proposed algorithm has better performance to improve coding efficiency.

• The Journal of Korean Institute of Information Technology
• /
• v.15 no.12
• /
• pp.165-171
• /
• 2017

This paper proposes a method of support-generation using morphological image processing instead of geometric calculations. The geometric computational cost is dependent on the shape, but our method is independent on the shape. For obtaining the external support area for extrusion shape, we represents morphological operations between two sliced layer images and shows results of each operation stages. Internal support area is evaluated from erosion and opening operations with the sliced-layer image. In these support areas, the supporter image is generated using the designed support structures. Also, we made a DLP printer and the STL model included supporter-structure is printed by the DLP printer. We confirmed the necessity of support-generation method with the support structures individually dependent on materials by looking at the printed results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.10
• /
• pp.5015-5038
• /
• 2018

In this paper, a novel spectral-spatial joint sparse representation algorithm for hyperspectral image classification is proposed based on multi-layer superpixels in various scales. Superpixels of various scales can provide complete yet redundant correlated information of the class attribute for test pixels. Therefore, we design a joint sparse model for a test pixel by sampling similar pixels from its corresponding superpixels combinations. Firstly, multi-layer superpixels are extracted on the false color image of the HSI data by principal components analysis model. Secondly, a group of discriminative sampling pixels are exploited as reconstruction matrix of test pixel which can be jointly represented by the structured dictionary and recovered sparse coefficients. Thirdly, the orthogonal matching pursuit strategy is employed for estimating sparse vector for the test pixel. In each iteration, the approximation can be computed from the dictionary and corresponding sparse vector. Finally, the class label of test pixel can be directly determined with minimum reconstruction error between the reconstruction matrix and its approximation. The advantages of this algorithm lie in the development of complete neighborhood and homogeneous pixels to share a common sparsity pattern, and it is able to achieve more flexible joint sparse coding of spectral-spatial information. Experimental results on three real hyperspectral datasets show that the proposed joint sparse model can achieve better performance than a series of excellent sparse classification methods and superpixels-based classification methods.

• Journal of Korean Society for Geospatial Information Science
• /
• v.15 no.4
• /
• pp.31-39
• /
• 2007

As the availability of images from airborne digital camera with high resolution is expanded, a lot of concern are shown about the production of orthoimage and digital map. This study presents the method of updating digital map using orthoimage from airborne digital camera image. Images were georectified using GPS surveying data. For the generation of orthoimage, Lidar DEM was used. The absolute positional accuracy of orthoimage was evaluated using GPS surveying data. And that of the building layer of digital map was estimated using the existed digital map at the scale of 1:1,000. The absolute positional accuracy of orthoimage was as followed: RMSE in X and Y were ${\pm}0.076m$ and ${\pm}0.294m$. The RMSE of the building layer were ${\pm}0.250m$ and ${\pm}0.210m$ in X and Y directions, respectively. The RMSE of the digital map using orthoimage from Aerial Digital Camera image fell within allowable error range established by NGII. Consequently, updating digital map using orthoimage from Aerial Digital Camera image can be applied to various fields including the construction of the framework data and the GIS of local government.

• Journal of Korea Technical Association of The Pulp and Paper Industry
• /
• v.44 no.4
• /
• pp.16-24
• /
• 2012

To characterize the coating structure, diverse methods such as mercury intrusion, nitrogen adsorption and oil absorption methods have been developed and widely employed. These indirect techniques, however, have some limitation to explain the actual coating structure. Recently microscopic observation methods have been tried for analyzing structural characteristics of coating layers. Preparation of the undamaged cross section of a coating layer is essential for obtaining high quality image for analysis. In this study, distortion-free cross-section of the coating layer was prepared using a grinding and polishing technique. The coated paper was embedded in epoxy resin and cured. After curing the resin block it was ground with abrasive papers and then polished with diamond particle suspension and nylon cloth. Polished coating layer was sufficient enough to obtain undamaged cross sectional images with scanning electron microscope under backscattered electron image mode. In addition, the SEM images allowed distinction of the coating layer components. Also S/B latex film formed between pigment particles was visualized by osmium tetroxide staining. Pore size distribution and pore orientation were evaluated by image analysis from SEM cross-sectional images.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1995.06a
• /
• pp.97-100
• /
• 1995

In transmission of image sequences over ATM networks, the packet loss problem and channel sharing efficiency are important. As a possible solution two-layer video coding methods have been proposed. These methods transmit video information over the network with different levels of protection with respect to packets loss. In this paper, a two-layer coding method using pyramid structure is proposed and several realizations of two-layer video coding methods are presented and their performances are compared.

• Archives of design research
• /
• v.17 no.2
• /
• pp.471-478
• /
• 2004

The recent aesthetical argument has been focusing on the interrelation between art and media. Now media is a kind of creative activity and dynamic factor which comprises aesthetic perception in itself compared to that of the past linear thingking era that was just a means of proving presence. Norbert Bolz, a leading media-aesthetician, puts the emphasis on the transformation of image, not the its dupe. He urges that this artistic form does not deliver the image of object, but means the objectless image itself. Therefore creativity is caused by an effect of a optional technic in the stage of technological environment where is under the control of the non-linear thingking system. In this context, this essay investigates to confirm the above mentioned aestheical views by analyzing the computer based layer as an detailed example, and to illuminate the designers creative activity where is deemed to be under its radical influence. Therefore it can be dedicated to the extension of the theoretical background in design, bridging between two practices accor ding to the Bolz's notion that the new aesthetical concept should be accompanied by ' the Design Science ', that is centered on the new media.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.44 no.6
• /
• pp.477-483
• /
• 2016

Velocity profiles of laminar, transition and turbulent boundary layers were investigated by using Particle Image Velocimetry(PIV) measurements on the flat plate at Mach 2.96. The Schlieren visualization and PIV measurements are also used to confirm whether the oblique shock wave generated from the leading edge affects the flow field over the flat plate. The laminar velocity profile measured from the experiment was well matched with the compressible Blasius solution. The velocity profile of the transition boundary layer was well correlated with the theoretical turbulent velocity profile from near the wall and the transition began from Re = $1.41{\times}106$. For the turbulent boundary layer, considering compressibility effects, the Van Driest-transformed velocity satisfies the incompressible log-law. It is found that the log region is extended farther in the wall-normal direction compared to the log region in incompressible boundary layer.