• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.8A
• /
• pp.779-785
• /
• 2002

This paper presents a phase-based stereo matching algorithm in order to efficiently extract 3-dimensional objects from two 2D images. Conventional phase-based methods, especially using windowed Fourier phases, inherit good properties in the case of hierarchical approaches, because they basically use a multi-resolution phase map. On the contrary, their computational costs are very heavy. Therefore, a fast hierarchical approach, using multi-resolution phase-based strategy and reducing the redundancy of phase calculations, is proposed in this pare. In addition, a structural matching algorithm on the phase domain is adopted to improve the matching quality. In experimental results, it is shown that the computation loads are considerably reduced about 8 times and stable outputs are obtained.

• Journal of Broadcast Engineering
• /
• v.24 no.6
• /
• pp.1064-1071
• /
• 2019

Multiview stereo matching algorithm is used to reconstruct 3D shape from a set of 2D images. Conventional multiview stereo algorithms have been implemented on high-performance hardware due to the heavy complexity that contains a large number of calculations in each step. However, as the performance of mobile graphics processors has recently increased rapidly, complex computer vision algorithms can now be implemented on mobile devices like a smartphone and an embedded board. In this paper we parallelize an multiview stereo algorithm using OpenCL on mobile GPU and provide various optimization techniques on the embedded hardware with limited resource.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.28 no.2
• /
• pp.105-112
• /
• 2017

In order to satisfy the diffusion of multimedia service in mobile communication and the demand for high-speed communication, it is essential to modify and improve high efficiency, wideband and nonlinear characteristic of multiband power amplifier. This research is designed to implement a single-stub matching circuit as a 2nd harmonic one that meets conditions of the Class-J power amplifier. Low characteristic impedance of the single-stub line is necessary to suit conditions of wideband Class-J. This research uses ceramic substrates having high permittivity to implement the single-stub line with low characteristic impedance, which eventually results in an amplifier satisfying the output impedance terms of Class-J in wideband frequency range. This result attributes to use of GaN HEMT packaged with a 2nd harmonic matching circuit and external fundamental circuit. The measurement results of the Class-J amplifier confirms the following characteristics: more than output power of 50 W(47 dBm) in bandwidth of 1.8~2.7 GHz(0.9GHz), maximum drain efficiency of 72.6 %, and maximum PAE characteristic of 66.5 %.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.1
• /
• pp.107-115
• /
• 2014

The reduced-bit transform based bit-plane matching algorithm (BPM) can obtain the block matching result through its simple calculation and hardware design compared to the conventional block matching algorithms (BMAs), but the block matching accuracy of BPMs is somewhat low. In this paper, reduced-bit transform based sum of the absolute difference (R-SAD) is proposed to improve the block matching accuracy in comparison with the conventional BPMs and it is shown that the matching process can be obtained using the logical operations. Firstly, this method transforms the current and the reference images into their respective 2-bit images and then a truth table is obtained from the relation between input and output 2-bit images. Next, a truth table is simplified by Karnaugh map and the absolute difference is calculated by using simple logical operations. Finally, the simulation results show that the proposed R-SAD can obtain higher accuracy in block matching results compared to the conventional BPMs through the PSNR analysis in the motion compensation experiments.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.2
• /
• pp.303-312
• /
• 2009

In this paper, we design and fabricate a disk sector antenna fed by CPW fur UWB communications. Also, we insert a rectangular slit on the arc-edge of the disk sector antenna. Then, the antenna has the directivity of E-면. In order to design the antenna, the input impedance is matched with the feed line of $50{\Omega}$ as varying the physical antenna parameters, which are the radius, the flare angle of disk sector, the length of ground, and the length of ground comer near by feed tine. Dimension of the antenna designed for UWB communication is $72mm{\times}26mm$ and bandwidth through computer simulation is $3{\sim}13GHz$. From the measured results, the bandwidth is $1.98{\sim}11GHz$. Return loss and gain of the fabricated antenna are -50.38dB, 1.34dBi at 3.5GHz, -12.27dB, 3.35dBi at 5.5GHz, -23.2dB, 3.8dBi at 8GHz and -16.17dB, 5.2dBi at 10GHz, respectively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.42 no.4
• /
• pp.798-805
• /
• 2017

This paper presents a low profile multi circular ring with different radius, patch antenna with modified feed line and slotted ground. The size of the antenna is $15{\times}12mm^2$, having electrical dimensions of $0.14{\lambda}{\times}0.12{\lambda}$ (at lower initial frequency) and footprints of $180mm^2$. The proposed antenna covers 3.1 to 12.3 GHz, reflection coefficient up to -38 dB with Bandwidth ratio of 4.13:1 and fractional Bandwidth of 122%. Bandwidth dimension ratio has been calculated (which is 6246) as a proof for compact size. Thin slots introduced on the feed line provide good impedance matching for whole frequency band. Numerical simulations of the proposed antenna are reported in terms of reflection coefficient ${\leq}-10dB$, vswr 2:1, radiation pattern and group delay (ns). The proposed antenna has advantage of very small size along with better impedance match that provides a practical approach to realize it for BW enhancement and UWB applications.

• The Proceeding of the Korean Institute of Electromagnetic Engineering and Science
• /
• v.6 no.2
• /
• pp.19-26
• /
• 1995

The design of broad-band, thin layer electromagnetic wave absorbers usually employs magnetic materials. The common absorber consists of a magnetic material layer bonded to a metal plate at the back. For such structure, a theoretical expression involving six parameters defines the allowable material shape and values of electromagnetic parameters that will produce zero reflection or more practically more than 99% absorption (20dB attenuation). The Secant Method was employed to compute the matching boundary values corresponding to 20dB attenuation. In addition, a search algorithm coupled with the Secant Method was developed to obtain allowable range of material parameters for the design of wave absorbers. These numerical methodes applicable to any six parameters effectively yielded accurate solutions in the least iterations.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.16 no.2 s.93
• /
• pp.228-234
• /
• 2005

A magnetic composite as noise absorber of quasi-microwave band was developed. The Fe-Si-Al alloy powder were forged by attrition mill to get flaky shape. The magnetic composite sheet was fabricated in which powders are dispersed in polymer and aligned in the direction perpendicular to electromagnetic wave propagation. The permittivity of magnetic composite is increased as forging time increasing, while the permeability is decreased slightly. The maximum attenuation peak of reflection loss is shifted to lower fiequency range as milling time increasing, and the value of maximum attenuation peak is to get smaller gradually. From these result, we could designed a noise absorber sheet (t=1.0 mm) for quasi-microwave band, which is impedance matched at 1.4 GHz with respect to -8.2 dB reflection 1055.

• Journal of Digital Convergence
• /
• v.10 no.10
• /
• pp.283-288
• /
• 2012

3D is the one of the current world in the spotlight as part of the future earnings of the business sector. Existing flat 2D and stereoscopic 3D to change the 3D shape and texture make walking along the dimension of the real world and the virtual reality world by making it feel contemporary reality of coexistence good show. 3D for the interest of the people has been spreading throughout the movie which is based on a 3D Avata. 3D TV market of the current conglomerate of changes in the market pioneer in the 3D market, further leap into the era of the upgrade was. At the same time, however, the modern man of the world, if becoming a necessity in the smartphone craze new innovation in the IT market mobile phone market and also has made. A small computer called a smartphone enough, the ripple velocity and the aftermath of the innovation of the telephone, the Internet as much as to leave many issues. Smartphone smart phone is a mobile phone that can be several functions. The current iPhone, Android. In addition to a large number of Windows Phone smartphones are released. Above the overall prospects of the future and a business service model for 3D facial expression as input avatar virtual 3D character on camera on your smartphone camera to recognize a user's emotional expressions on the face of the person is able to synthetic synthesized avatars in real-time to other mobile phone users matching, transmission, and be able to communicate in real-time sensibility fused video communication services to the development of applications.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.613-615
• /
• 2001

본 논문은 동영상 내의 얼굴을 특정인 얼굴로 자동 변환 및 정합하는 기술을 제안한다. 얼굴에 나타난 동작이나 표정은 높은 자유도로 인하여 기존에 사용되어온 2차원적이고 고정된 물체 위주의 동영상 정합 기술로는 자연스러운 결과물을 얻기가 어렵다. 본 논문에서는 입력 받은 정면 유사방향의 사진으로부터 3차원 얼굴 모델을 복원한다. 각 프레임에 등장한 얼굴의 3차원 방향을 추출하여 복원한 3차원 얼굴 모델에 적용한 후 대체할 얼굴 영역에 저합시킨다. 정합 과정 시 비디오 프레임 내의 조명효과와 얼굴색 등을 분석하고 3차원 얼굴 모델에 블렌딩하여 비디오 프레임과 자연스럽게 정합할 수 있도록 한다.