• ETRI Journal
• /
• 제44권2호
• /
• pp.327-338
• /
• 2022

Classification models for human action recognition require robust features and large training sets for good generalization. However, data augmentation methods are employed for imbalanced training sets to achieve higher accuracy. These samples generated using data augmentation only reflect existing samples within the training set, their feature representations are less diverse and hence, contribute to less precise classification. This paper presents new data augmentation and action representation approaches to grow training sets. The proposed approach is based on two fundamental concepts: virtual video generation for augmentation and representation of the action videos through robust features. Virtual videos are generated from the motion history templates of action videos, which are convolved using a convolutional neural network, to generate deep features. Furthermore, by observing an objective function of the genetic algorithm, the spatiotemporal features of different samples are combined, to generate the representations of the virtual videos and then classified through an extreme learning machine classifier on MuHAVi-Uncut, iXMAS, and IAVID-1 datasets.

• 한국항해항만학회:학술대회논문집
• /
• 한국항해항만학회 2006년도 International Symposium on GPS/GNSS Vol.2
• /
• pp.33-37
• /
• 2006

Since 1993, the civil aviation community through RTCA (Radio Technical Commission for Aeronautics) and the ICAO (International Civil Air Navigation Organization) have been working on the definition of GNSS augmentation systems that will provide improved levels of accuracy and integrity. These augmentation systems have been classified into three distinct groups: Aircraft Based Augmentation Systems (ABAS), Space Based Augmentation Systems (SBAS) and Ground Based Augmentation Systems (GBAS). The last one is an implemented system to support Air Navigation in CAT-I approaching operation. It consists of three primary subsystems: the GNSS Satellite subsystem that produces the ranging signals and navigation messages; the GBAS ground subsystem, which uses two or more GNSS receivers. It collects pseudo ranges for all GNSS satellites in view and computes and broadcasts differential corrections and integrity-related information; the Aircraft subsystem. Within the area of coverage of the ground station, aircraft subsystems may use the broadcast corrections to compute their own measurements in line with the differential principle. After selection of the desired FAS for the landing runway, the differentially corrected position is used to generate navigation guidance signals. Those are lateral and vertical deviations as well as distance to the threshold crossing point of the selected FAS and integrity flags. The Department of Applied Science in Naples has create for its study a virtual GBAS Ground station. Starting from three GPS double frequency receivers, we collect data of 24h measures session and in post processing we generate the GC (GBAS Correction). For this goal we use the software Pegasus V4.1 developed from EUROCONTROL. Generating the GC we have the possibility to study and monitor GBAS performance and integrity starting from a virtual functional architecture. The latter allows us to collect data without the necessity to found us authorization for the access to restricted area in airport where there is one GBAS installation.

• 한국정밀공학회지
• /
• 제20권12호
• /
• pp.73-79
• /
• 2003

Recently, the tole-robot operations to an unstructured environment have been widely researched. The human's interaction with the tole-robot system can be used to improve robot operation and performance for an unknown environment. The exact modeling based on real environment is fundamental and important process for this interaction. In this paper, we propose an extrinsic parameter calibration and data augmentation method that only uses a circular jig in the hand-eye laser virtual environment. Compared to other methods, easier estimation and overlay can be done by this algorithm. Experimental results using synthetic graphic demonstrate the usefulness of the proposed algorithm.

• 한국정보통신학회:학술대회논문집
• /
• 한국해양정보통신학회 2011년도 춘계학술대회
• /
• pp.369-371
• /
• 2011

증강현실 기술을 이용하여 데이터 가시화을 통하여 사용자들은 새로운 경험을 할 수 있기 때문에 교육용 어플리케이션에 적합한 미디어라고 할 수 있다. 본 논문에서는 화학실험에서의 위험성을 배제하고 몰입형 실험을 하기 위하여 증강현실을 이용한 실험 시스템이다. 가상 화학 실험 시스템은 사용자의 손을 활용하여 가상의 객체와 상호작용을 이룰 수 있는 인터페이스를 통해, 기존 증강현실의 몰입감 저하의 문제점인 마커를 사용하지 않고, 3D 객체를 컨트롤 하는 기법을 사용했다. 시스템에 대한 몰입감을 극대화하기 위해서 협업이 가능한 가상 화학 실험 시스템을 제안했다.

• 산업경영시스템학회지
• /
• 제25권6호
• /
• pp.12-16
• /
• 2002

Virtual reality (VR) opened the possibility of moving beyond the 2-D world of conventional desktop systems to an immersive, multi-sensory environment generated by computer graphics. Haptic feedback (Haptics) is now starting to get recognition and use in intensive applications for manipulation, while smell and taste feedback are at the stage of early research. This paper reviews some current works on increased haptic feedback augmentation in virtual reality applications.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2006년도 하계종합학술대회
• /
• pp.685-686
• /
• 2006

Recently, various AR-based product design methodologies have been introduced. In this paper, we propose technologies for enhancing robust augmentation and immersive realization of virtual objects. A robust augmentation technology is developed for various lighting conditions and a partial solution is proposed for the hand occlusion problem that occurs when the virtual objects overlay the user' hands. It provides more immersive or natural images to the users. Finally, vibratory haptic cues by page motors as well as button clicking force feedback by modulating pneumatic pressures are proposed while interacting with virtual widgets. Also our system reduces gabs between modeling spaces and user spaces. An immersive game-phone model is selected to demonstrate that the users can control the direction of the car in the racing game by tilting a tangible object with the proposed augmented haptic and robust non-occluded visual feedback. The proposed methodologies will be contributed to the immersive realization of the conventional AR system.

• International Journal of Automotive Technology
• /
• 제6권5호
• /
• pp.545-554
• /
• 2005

Passive velocity field control (PVFC) was previously developed for fully mechanical systems, in which the motion task was specified by behaviors in terms of a velocity field and the closed-loop was passive with respect to the supply rate given by the environment input. However, the PVFC was only applied to a single manipulator. The proposed control law was derived geometrically and the geometric and robustness properties of the closed-loop system were also analyzed. In this paper, we propose a virtual passivity-based algorithm to apply decentralized control to multiple 3­wheeled mobile robotic systems whose subsystems are under nonholonomic constraints and convey a common rigid object in a horizontal plain. Moreover, it is shown that multiple robot systems ensure stability and the velocities of augmented systems converge to a scaled multiple of each desired velocity field for cooperative mobile robot systems. Finally, the application of proposed virtual passivity-based decentralized algorithm via system augmentation is applied to trace a circle and the simulation results is presented in order to show effectiveness for the decentralized control algorithm proposed in this research.

• 방송공학회논문지
• /
• 제25권2호
• /
• pp.275-278
• /
• 2020

본 논문에서는 가상 데이터 생성을 통한 딥러닝 기반 문자인식 시스템을 제안한다. 지도학습에서 가장 큰 비중을 차지하는 학습 데이터를 확보하기 위하여 가상 데이터를 생성하였다. 또한 가상 데이터를 생성 후 증강 파라미터를 이용하여, 실제 다양한 데이터에 대응하기 위해서 데이터 일반화를 하였다. 최종적으로 학습 데이터 구성은 증강 파라미터와 폰트 인자에 다양한 값을 대입하여 데이터를 생성하였다. 문자인식 성능을 측정하기 위한 테스트 데이터는 실제 촬영된 이미지 데이터에서 문자영역을 크롭하여 구성하였다. 테스트 데이터는 실제환경에서 발생할 수 있는 이미지 왜곡을 고려하여 데이터 증강하였다. 딥러닝 알고리즘은 실시간 검출에 용이한 YOLO v3를 사용하였으며, 추론결과는 후처리를 통하여 최종 검출결과를 출력한다.

• Journal of Mechanical Science and Technology
• /
• 제19권9호
• /
• pp.1720-1730
• /
• 2005

Automatic guided vehicle in the factory has an important role to advance the flexible manufacturing system. In this paper, we propose a novel object-transportation control algorithm of cooperative AGV systems to apply decentralized control to multiple AGV systems. Each AGV system is under nonholonomic constraints and conveys a common object-transportation in a horizontal plain. Moreover it is shown that cooperative robot systems ensure stability and the velocities of augmented systems convergence to a scaled multiple of each desired velocity field for cooperative AGV systems. Finally, the application of proposed virtual passivity-based decentralized control algorithm via system augmentation is applied to trace a circle. Finally, the simulation and experimental results for the object-transportation by two AGV systems illustrates the validity of the proposed virtual-passivity decentralized control algorithm.

• KEPCO Journal on Electric Power and Energy
• /
• 제7권2호
• /
• pp.323-328
• /
• 2021

Diagnostic models are required. Data augmentation is one of the best ways to improve deep learning performance. Traditional augmentation techniques that modify image brightness or spatial information are difficult to achieve great results. To overcome this, a generative adversarial network (GAN) technology that generates virtual data to increase deep learning performance has emerged. GAN can create realistic-looking fake images by competitive learning two networks, a generator that creates fakes and a discriminator that determines whether images are real or fake made by the generator. GAN is being used in computer vision, IT solutions, and medical imaging fields. It is essential to secure additional learning data to advance deep learning-based fault diagnosis solutions in the power industry where facilities are strictly maintained more than other industries. In this paper, we propose a method for generating power facility images using GAN and a strategy for improving performance when only used a small amount of data. Finally, we analyze the performance of the augmented image to see if it could be utilized for the deep learning-based diagnosis system or not.