• 대한임베디드공학회논문지
• /
• 제16권6호
• /
• pp.323-330
• /
• 2021

Visual object tracking is a challenging area of study in the field of computer vision due to many difficult problems, including a fast variation of target shape, occlusion, and arbitrary ground truth object designation. In this paper, we focus on the reinforced feature of the dynamic search area to get better performance than conventional discriminative model prediction trackers on the condition when the accuracy deteriorates since low feature discrimination. We propose a reinforced input feature method shown like the spotlight effect on the dynamic search area of the target tracking. This method can be used to improve performances for deep learning based discriminative model prediction tracker, also various types of trackers which are used to infer the center of the target based on the visual object tracking. The proposed method shows the improved tracking performance than the baseline trackers, achieving a relative gain of 38% quantitative improvement from 0.433 to 0.601 F-score at the visual object tracking evaluation.

• 한국군사과학기술학회지
• /
• 제8권2호
• /
• pp.49-57
• /
• 2005

The target localization using the line arrays buried in the seabed is a difficult problem due to the complex sea bottom characteristics and need to compensate the wave propagation effect to localize the target accurately Sound speed mismatch in the seabed causes a bias in the target bearing estimation and induces the localization error. In this paper we describe a target localization method with improved accuracy of target bearing and localization by calibration the sound speed in the seabed. The proposed algorithm is verified through the ocean data.

• 한국멀티미디어학회논문지
• /
• 제23권2호
• /
• pp.186-203
• /
• 2020

Image-based Virtual Try-On Systems are among the most potential solution for virtual fitting which tries on a target clothes into a model person image and thus have attracted considerable research efforts. In many cases, current solutions for those fails in achieving naturally looking virtual fitted image where a target clothes is transferred into the body area of a model person of any shape and pose while keeping clothes context like texture, text, logo without distortion and artifacts. In this paper, we propose a new improved image-based virtual try-on network system based on keypoints, which we name as KP-VTON. The proposed KP-VTON first detects keypoints in the target clothes and reliably predicts keypoints in the clothes of a model person image by utilizing a dense human pose estimation. Then, through TPS transformation calculated by utilizing the keypoints as control points, the warped target clothes image, which is matched into the body area for wearing the target clothes, is obtained. Finally, a new try-on module adopting Attention U-Net is applied to handle more detailed synthesis of virtual fitted image. Extensive experiments on a well-known dataset show that the proposed KP-VTON performs better the state-of-the-art virtual try-on systems.

• 한국정보통신학회논문지
• /
• 제21권12호
• /
• pp.2371-2379
• /
• 2017

넓은 탐색영역에서 활동하는 드론에서 신속한 처치를 요하는 응급환자의 탐색, 신속한 경보와 대응을 요하는 자연재해의 감시와 같은 응용 분야에서 목표물 파악의 시간(time), 즉 신속성의 문제는 매우 중요한 문제가 된다. 드론의 실제 운영에 있어서 목표물을 파악하는 시간은 탐색 영역을 효율적으로 탐색하기 위한 탐색 알고리즘 및 드론 간의 협업과 매우 연관성이 깊다. 따라서 본 연구에서는 드론을 이용한 목표물 탐색에 있어서 신속성의 문제를 해결하기 위하여, 고도를 달리하는 드론들의 협력에 의한 확률기반 목표물 탐색 방법을 제안한다. 특히 제안한 방법은 고(高)고도 드론이 우선 탐색을 실시하고, 탐색 결과를 저(低)고도 드론에 전달하여 보다 정밀한 탐색을 함으로써 탐색 시간을 줄이고 목표물 발견의 확률을 높이는 방법이다. 시뮬레이션을 통하여 제안된 방법의 성능을 분석한다.

• 한국군사과학기술학회지
• /
• 제12권4호
• /
• pp.500-507
• /
• 2009

This paper presents a novel method which can detect a target in IR image for active protection system. The target in IR image for the active protection system is small, moreover it moves with enormous speed. The proposed algorithm is comprised of robust clutter rejection methods and target optimized detection algorithms for small target, and an advanced method of selecting a final target position in target area, it can work in some milliseconds. The proposed algorithm provides the active protective system with more correct positions than those of radar, so that helps the active protection system can defense all threats with the utmost precision.

• 융합보안논문지
• /
• 제16권5호
• /
• pp.65-71
• /
• 2016

최근 무선 센서 네트워크는 다양한 분야에 적용되고 있다. 무선 센서 네트워크는 안정적인 네트워크 설계 뿐만 아니라 보안이나, 군 그리고 병원의 응급 처리에도 적용되고 있다. 이러한 다양한 응용 중에서 어떤 침입자나 위기 상황이 발생했을 경우 이를 신속히 위치를 추적하는 방법이 아주 필수적인 연구 분야가 되고 있다. 이러한 방법을 측위라고 정의하고, 센서 노드의 전파범위를 기반으로 측위를 효율적으로 처리하는 기법을 제안한다. 또한 측위를 위하여 필수적인 효율적인 클러스터링 방법과 알고리즘을 제안한다.

• 제어로봇시스템학회논문지
• /
• 제9권12호
• /
• pp.1033-1041
• /
• 2003

A new visual tracking scheme is proposed for a mobile robot that tracks a moving object in 3D space in real time. Visual tracking is to control a mobile robot to keep a moving target at the center of input image at all time. We made it possible by simplifying the relationship between the 2D image frame captured by a single camera and the 3D workspace frame. To precisely calculate the input vector (orientation and distance) of the mobile robot, the speed vector of the target is determined by eliminating the speed component caused by the camera motion from the speed vector appeared in the input image. The problem of temporary disappearance of the target form the input image is solved by selecting the searching area based on the linear prediction of target motion. The experimental results have shown that the proposed scheme can make a mobile robot successfully follow a moving target in real time.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2000년도 하계학술대회 논문집 C
• /
• pp.1485-1487
• /
• 2000

A variety of target configurations in sputtering process have been proposed to deposit various structures of thin film alloys and compound films. In this study, we presented the comparative experimental results regarding to the characterization of properties of Cu-Ni thin films deposited by using a magnetron co-sputtering method, as a function of target configurations; one is using a single target with varying the area of Ni chips attached on the Cu target and another is using a dual-type target with two targets of Ni and Cu separated each other. Structural(d-spacing, crystal orientation, crystallite size, cross-sectional morphology) and electrical(resistivity) properties of deposited films are characterized and compared as a function of target configurations as well as deposition conditions.

• 한국군사과학기술학회지
• /
• 제13권3호
• /
• pp.445-452
• /
• 2010

Adequate segmentation of target objects from the background plays an important role for the performance of automatic target recognition(ATR) system. This paper presents a new segmentation algorithm using fuzzy thresholding to extract a target. The proposed algorithm consists of two steps. In the first step, the region of interest(ROI) including the target can be automatically selected by the proposed robust method based on the frame difference of each image sensor. In the second step, fuzzy thresholding with a proposed membership function is performed within the only ROI selected in the first step. The proposed membership function is based on the similarity of intensity and the adjacency of target area on each image. Experimental results applied to real CCD/IR images show a good performance and the proposed algorithm is expected to enhance the performance of ATR system using multi-sensors.

• 한국환경복원기술학회지
• /
• 제21권3호
• /
• pp.57-65
• /
• 2018

The study is conducted to establish the National Protected Areas Integrated System(NPAIS) which includes National Protected Areas(NPAs) and other conservation measures in terms of effective ways for biodiversity conservation. Additionally, it is carried out to evaluate progress toward quantitative expansion goals in Aichi biodiversity Target 11. The NPAIS consists of NPAs and other effective area-based conservation measures(OECMs). There are two different types of OECMs. OECMs type I, including water-source protection Areas(WPA), riparian zones(RZ), fishery-resource protection zones (FPZ), and urban natural park zones(UNPZ), is a potential protected area which is recommended to be incorporated into the NPAs for effective management. OECMs type II means development restriction zones(DRZ), natural recreation forests(NRF), and buffer zones for Korea national arboretum(BKNA). As a result of evaluating the quantitative expansion goals of the NPAIS, terrestrial and inland water protected areas exceed 17% of the quantitative goal in Aichi biodiversity Target 11. The larger the area of individual OECMs and the lower the degree of overlap with NPAs, the higher the contribution of them to the terrestrial and inland water protected areas. DRZ contributes to enlarge more than 3% of quantitative expansion. And RZ and NRF contribute more than 1%. The marine protected areas are expanded by $1,425km^2$ through FPZ, but the expanded area is very small as comparing with the total marine area. It adds only 0.321% to the quantitative expansion. The rest of OECMs also has very poor quantitative expansion contributions in the marine area. Consequently, the NPAIS is effective for quantitative expansion of land areas, but not for marine areas.