• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.1
• /
• pp.29-34
• /
• 2022

We propose a novel network architecture and build dataset for recognizing clickable objects on mobile device screens. The data was collected based on clickable objects on the mobile device screen that have numerous resolution, and a total of 24,937 annotation data were subdivided into seven categories: text, edit text, image, button, region, status bar, and navigation bar. We use the Deconvolution Single Shot Detector as a baseline, the backbone network with Squeeze-and-Excitation blocks, the Single Shot Detector layer structure to derive inference results and the Feature pyramid networks structure. Also we efficiently extract features by changing the input resolution of the existing 1:1 ratio of the network to a 1:2 ratio similar to the mobile device screen. As a result of experimenting with the dataset we have built, the mean average precision was improved by up to 101% compared to baseline.

• Journal of the Optical Society of Korea
• /
• v.14 no.4
• /
• pp.409-414
• /
• 2010

This paper describes the surface form measurement of a spherical smooth surface by using single shot off-axis Fizeau interferometry. The demodulated phase map is obtained and unwrapped to remove the $2\pi$ ambiguity. The unwrapped phase map is converted to height and the 3D surface height of the surface object is reconstructed. The results extracted from the single shot off-axis geometry are compared with the results extracted from four-frame phase shifting in-line interferometry, and the results are in excellent agreement.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.10a
• /
• pp.625-627
• /
• 2018

브레이크 패드 제작 공정에서 문자 인식은 사람이 직접 인식하거나 컴퓨터 비전 기술의 역할이었다. 하지만 사람의 인식 오류나 잉크가 번진 문자같은 새로운 형태의 문자를 인식하지 못하는 비전 기술의 단점 등 많은 한계가 존재했다. 본 논문에서는 C/CUDA로 설계한 Single Shot Multibox Detector 기반 Inference Program 을 통해 더 정확한 문자인식 결과를 제시하고, CUDA를 이용한 향상된 연산속도를 통해 실시간 문자 인식이 가능하도록 하였다. 문자 인식 정확도는 약 96.6%로 기존 비전 기술보다 더 뛰어난 성능을 보였다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.9-12
• /
• 2019

지난 수 년 동안 계속해서 일반 실상 카메라를 이용한 영상분석기술에 대한 연구가 활발히 진행되고 있다. 최근에는 딥러닝 기술을 적용한 지능형 영상분석기술로 발전해 왔으며 국방기지방호, CCTV, 사용자 얼굴인식, 머신비전, 자동차, 드론 산업이 활성화되면서 많은 시너지를 효과를 일으키고 있다. 그러나 어두운 밤과 안개, 날씨, 연기 등 다양한 여건에서 따라서 카메라의 영상분석 정확성 감소와 오류가 수반될 수 있으며 일반적으로 딥러닝 기술을 활용하기 위해서는 고사양의 GPU를 필요로 하기 때문에 다른 추가적인 시스템이 요구된다. 이에 본 연구에서는 열적외선 영상의 객체 검출에 적용하기 위해 SSD(Single Shot MultiBox Detector) 기반의 경량적인 MobilNet 네트워크로 재구성하여, 모바일 기기 등 낮은 사양의 낮은 임베디드 시스템에서도 활용 할 수 있는 방법을 제안한다. 모의 실험결과 제안된 방식의 모델은 적외선 열화상 카메라에서 객체검출과 학습시간이 줄어든 것을 확인 할 수 있었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.358-361
• /
• 2021

In order to improve the accuracy of small target detection more effectively, this paper proposes an improved single shot detector (SSD) target detection and recognition method based on cspdarknet53, which introduces lightweight ECA attention mechanism and Feature Pyramid Network (FPN). First, the original SSD backbone network is replaced with cspdarknet53 to enhance the learning ability of the network. Then, a lightweight ECA attention mechanism is added to the basic convolution block to optimize the network. Finally, FPN is used to gradually fuse the multi-scale feature maps used for detection in the SSD from the deep to the shallow layers of the network to improve the positioning accuracy and classification accuracy of the network. Experiments show that the proposed target detection algorithm has better detection accuracy, and it improves the detection accuracy especially for small targets.

• Bulletin of the Korean Space Science Society
• /
• 2008.10a
• /
• pp.25.3-26
• /
• 2008

The direct way to the moon is to start from the parking orbit by using impulsive thruster In previous domestic research, the direct way has been studied by using a single impulsive shot. However, when a single impulsive shot occurs to go into a Translunar orbit, gravity losses occur because thruster is not impulsive shot but the finite burns and it causes the gravity losses. To make up for the weak point of a single impulsive shot, this paper divides TLI (Trans Lunar Injection) into several small burns. Therefore, departure loop trajectory and the Translunar trajectory. This method is useful not only to reduce the gravity losses but also to check the condition of satellite. By using this method, this paper demostrates the optimized trajectory from Earth parking orbit to lunar mission orbit which minimizes the fuel, and the SNOPT (Sparse Nonlinear OPTimizer software) is used to find optimal solution. Also, this paper provides lunar mission profile which includes the mission schedule when TLI, LOI (Lunar Orbit Insertion) maneuvers occur, a mount of fuel when thruster is used and other mission parameters.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.5C
• /
• pp.514-522
• /
• 2003

Contrary to the previous methods, in this paper, we propose an enhanced shot change detection method using multiple features and binary decision tree. The previous methods usually used single feature and fixed threshold between consecutive frames. However, contents such as color, shape, background, and texture change simultaneously at shot change points in a video sequence. Therefore, in this paper, we detect the shot changes effectively using multiple features, which are supplementary each other, rather than using single feature. In order to classify the shot changes, we use binary classification tree. According to this classification result, we extract important features among the multiple features and obtain threshold value for each feature. We also perform the cross-validation and droop-case to verify the performance of our method. From an experimental result, it was revealed that the EI of our method performed average of 2% better than that of the conventional shot change detection methods.

• Journal of the Korean Society of Radiology
• /
• v.10 no.4
• /
• pp.279-284
• /
• 2016

Single-shot echo planar imaging(SS-EPI) is well established as high sensitivity for ischemic stroke. However, it is prone to susceptibility artifact in brain stem that diminish the image quality. single-shot turbo spin echo(SS-TSE) is a new DWI technique that can reduce susceptibility artifact. Thus, this research was conducted so as to reduce geometric distortion in brain stem by using single-shot turbo spin echo technique. Thirty patients without brain disease underwent diffusion MR on a 3T scanner with SS-EPI and SS-TSE. Obtained images with both sequences were analyzed for geometric distortion and error percentage as well. Image quality in terms of geometric distortion of SS-TSE were found to be significantly better than those for SS-EPI. And error percentage was considerably reduced for 2.4% of b0 image(from 11.1% to 8.7%), 1.2% of b1000 image(from 11.4% to 10.1%), respectively. In summary, diffusion MR using SS-TSE significantly reduce geometric distortion compared to SS-EPI in brain stem and may provide improved diagnostic performance.

• Korean Journal of Radiology
• /
• v.22 no.7
• /
• pp.1044-1053
• /
• 2021

Objective: Motion-corrected averaging with a single-shot technique was introduced for faster acquisition of late-gadolinium-enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging while free-breathing. We aimed to evaluate the image quality (IQ) of free-breathing motion-corrected single-shot LGE (moco-ss-LGE) in patients with hypertrophic cardiomyopathy (HCM). Materials and Methods: Between April and December 2019, 30 patients (23 men; median age, 48.5; interquartile range [IQR], 36.5-61.3) with HCM were prospectively enrolled. Breath-held single-shot LGE (bh-ss-LGE) and free-breathing moco-ss-LGE images were acquired in random order on a 3T MR system. Semi-quantitative IQ scores, contrast-to-noise ratios (CNRs), and quantitative size of myocardial scar were assessed on pairs of bh-ss-LGE and moco-ss-LGE. The mean ± standard deviation of the parameters was obtained. The results were compared using the Wilcoxon signed-rank test. Results: The moco-ss-LGE images had better IQ scores than the bh-ss-LGE images (4.55 ± 0.55 vs. 3.68 ± 0.45, p < 0.001). The CNR of the scar to the remote myocardium (34.46 ± 11.85 vs. 26.13 ± 10.04, p < 0.001), scar to left ventricle (LV) cavity (13.09 ± 7.95 vs. 9.84 ± 6.65, p = 0.030), and LV cavity to remote myocardium (33.12 ± 15.53 vs. 22.69 ± 11.27, p < 0.001) were consistently greater for moco-ss-LGE images than for bh-ss-LGE images. Measurements of scar size did not differ significantly between LGE pairs using the following three different quantification methods: 1) full width at half-maximum method; 23.84 ± 12.88% vs. 24.05 ± 12.81% (p = 0.820), 2) 6-standard deviation method, 15.14 ± 10.78% vs. 15.99 ± 10.99% (p = 0.186), and 3) 3-standard deviation method; 36.51 ± 17.60% vs. 37.50 ± 17.90% (p = 0.785). Conclusion: Motion-corrected averaging may allow for superior IQ and CNRs with free-breathing in single-shot LGE imaging, with a herald of free-breathing moco-ss-LGE as the scar imaging technique of choice for clinical practice.

• Journal of Ocean Engineering and Technology
• /
• v.15 no.4
• /
• pp.66-72
• /
• 2001

Recently the steel parts used at automobiles are required to be used under high stress more than ever before in need of the weight down. To achieve this requirement of a high strength steel, it must be necessary to decrease inclusion content and surface defect as like decarburization, surface roughness etc.. In this study, the surface conditions are measured to know the influence on fatigue properties by two cases of shot peening of two-stage shot peening and single-stage shot peening. And for this study, two kinds of spring steel (JISG4081SUP7, DIN 50CrV4) are used. This study shows the outstanding improvement of fatigue properties at the case of two-stage shot peening in the rotary bending fatigue test and this is assumed to be from on low stress condition, the 1st stage shot peening is not affected by nonmetallic inclusion under metal. it is possible that the 2nd stage shot peening increases the fatigue life and the high stress but that is affected by nonmetallic inclusion under metal. so far beeasily DIN50CrV4 have made high stress. But, results also show fatigue failures originated at inclusion near surface, and this inclusion type is turned out to be a alumina of high hardness.