• 한국멀티미디어학회논문지
• /
• 제16권12호
• /
• pp.1393-1402
• /
• 2013

Hand gesture is one of the most popular communication methods in everyday life. In human-computer interaction applications, hand gesture recognition provides a natural way of communication between humans and computers. There are mainly two methods of hand gesture recognition: glove-based method and vision-based method. In this paper, we propose a vision-based hand gesture recognition method using Kinect. By using the depth information is efficient and robust to achieve the hand detection process. The finger labeling makes the system achieve pose classification according to the finger name and the relationship between each fingers. It also make the classification more effective and accutate. Two kinds of gesture sets can be recognized by our system. According to the experiment, the average accuracy of American Sign Language(ASL) number gesture set is 94.33%, and that of general gestures set is 95.01%. Since our system runs in real-time and has a high recognition rate, we can embed it into various applications.

• 전자공학회논문지SC
• /
• 제48권6호
• /
• pp.91-98
• /
• 2011

본 논문에서는 수중로봇에 쓰일 수 있는 카메라 영상을 기반으로 하는 인공표식물의 인식 및 추종 기법을 제안한다. 문제를 풀기 위해 제안된 방법은 인식과 추종의 두 개의 단계로 이루어져 있으며 인식단계에서는 물체의 외형에 관한 특징을 분석한 후 비선형 최적화 알고리즘을 통하여 알맞은 목표물로 분류한다. 이 후 추종 단계에서는 분류된 목표물에서 색깔 히스토그램을 추출한 후 meanshift 추종 법을 이용하여 지속적으로 추종하는 방법을 택하였다. 히스토그램 매칭 시에는 Bhattacharyya 거리를 계산하는 방법을 이용하였다. 결과적으로 제안하는 접근법은 수중로봇의 영상처리 분야에 다음과 같은 공헌을 할 것으로 기대한다. 1) 제안하는 방법은 카메라의 움직임으로 생기는 물체의 자세변화나 크기 변화에도 강인하게 대처할 수 있으며 2) 카메라 센서를 통한 방법이므로 초음파 센서 등의 기기들에 비하여 가격 경쟁력이 우수하다. 3) 또한 본 논문에서는 일반적으로 많이 쓰이는 특징 점을 기반으로 한 방법이 탁도 변화에서는 형태를 기반으로 한 방법보다 열등할 수 있음을 실험을 통하여 보였다. 4) 마지막으로 제안된 방법의 성능을 기존의 방법들과 비교하여 수치적으로 검증해 보았다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권8호
• /
• pp.2749-2763
• /
• 2021

The desirable result of infrared (IR) and visible (VIS) image fusion should have textural details from VIS images and salient targets from IR images. However, detail information in the dark regions of VIS image has low contrast and blurry edges, resulting in performance degradation in image fusion. To resolve the troubles of fuzzy details in dark regions of VIS image fusion, we have proposed a method of reflectance estimation for IR and VIS image fusion. In order to maintain and enhance details in these dark regions, dark region approximation (DRA) is proposed to optimize the Retinex model. With the improved Retinex model based on DRA, quasi-Newton method is adopted to estimate the reflectance of a VIS image. The final fusion outcome is obtained by fusing the DRA-based reflectance of VIS image with IR image. Our method could simultaneously retain the low visibility details in VIS images and the high contrast targets in IR images. Experiment statistic shows that compared to some advanced approaches, the proposed method has superiority on detail preservation and visual quality.

• 한국항행학회논문지
• /
• 제10권2호
• /
• pp.113-120
• /
• 2006

본 논문에서는 로봇 매니퓰레이터가 기준 궤적을 추종할 수 있도록 영상기반 강인제어기를 제안하였다. 카메라로부터 획득된 목표물의 특징점을 원하는 특징점 좌표로 로봇이 이동할 수 있도록 기준 궤적을 생성하고, 제어 입력에 포함되는 로봇 동역학부의 파라미터 불확실성을 보상하기 위한 강인제어기를 설계하였다. 또한 자코비안에 불확실성이 존재하는 경우 이를 보상하기 위한 제어 입력을 갖는 영상기반 강인제어기를 제안하였다. 시스템의 안정도는 Lyapunov 안정도 판별법을 이용하여 검증하였다. 5-링크 2 자유도의 로봇 매니퓰레이터를 대상으로 제안된 제어기를 적용한 모의실험과 실험을 통하여 제어 성능을 입증하였다.

• 대한임베디드공학회논문지
• /
• 제18권3호
• /
• pp.89-100
• /
• 2023

Recently, on-device artificial intelligence (AI) solutions using mobile devices and embedded edge devices have emerged in various fields, such as computer vision, to address network traffic burdens, low-energy operations, and security problems. Although vision transformer deep learning models have outperformed conventional convolutional neural network (CNN) models in computer vision, they require more computations and parameters than CNN models. Thus, they are not directly applicable to embedded edge devices with limited hardware resources. Many researchers have proposed various model compression methods or lightweight architectures for vision transformers; however, there are only a few studies evaluating the effects of model compression techniques of vision transformers on performance. Regarding this problem, this paper presents a performance evaluation of vision transformers on embedded platforms. We investigated the behaviors of three vision transformers: DeiT, LeViT, and MobileViT. Each model performance was evaluated by accuracy and inference time on edge devices using the ImageNet dataset. We assessed the effects of the quantization method applied to the models on latency enhancement and accuracy degradation by profiling the proportion of response time occupied by major operations. In addition, we evaluated the performance of each model on GPU and EdgeTPU-based edge devices. In our experimental results, LeViT showed the best performance in CPU-based edge devices, and DeiT-small showed the highest performance improvement in GPU-based edge devices. In addition, only MobileViT models showed performance improvement on EdgeTPU. Summarizing the analysis results through profiling, the degree of performance improvement of each vision transformer model was highly dependent on the proportion of parts that could be optimized in the target edge device. In summary, to apply vision transformers to on-device AI solutions, either proper operation composition and optimizations specific to target edge devices must be considered.

• 대한인간공학회지
• /
• 제31권4호
• /
• pp.601-607
• /
• 2012

Objective: The aim of this study is to classify between intentional and natural blinks in vision based eye tracking system. Through implementing the classification method, we expect that the great eye tracking method will be designed which will perform well both navigation and selection interactions. Background: Currently, eye tracking is widely used in order to increase immersion and interest of user by supporting natural user interface. Even though conventional eye tracking system is well focused on navigation interaction by tracking pupil movement, there is no breakthrough selection interaction method. Method: To determine classification threshold between intentional and natural blinks, we performed experiment by capturing eye images including intentional and natural blinks from 12 subjects. By analyzing successive eye images, two features such as eye closed duration and pupil size variation after eye open were collected. Then, the classification threshold was determined by performing SVM(Support Vector Machine) training. Results: Experimental results showed that the average detection accuracy of intentional blinks was 97.4% in wearable eye tracking system environments. Also, the detecting accuracy in non-wearable camera environment was 92.9% on the basis of the above used SVM classifier. Conclusion: By combining two features using SVM, we could implement the accurate selection interaction method in vision based eye tracking system. Application: The results of this research might help to improve efficiency and usability of vision based eye tracking method by supporting reliable selection interaction scheme.

• 한국구조물진단유지관리공학회 논문집
• /
• 제27권2호
• /
• pp.58-66
• /
• 2023

이 연구에서는 진동법을 이용하여 케이블 장력을 추정하고 비접촉 방식으로 케이블의 변위 응답을 측정하기 위한 센서로 영상기반시스템을 적용하였다. 영상기반시스템에서 카메라는 대상 구조물과 카메라의 시야를 고려한 위치에 설치된다. 그러나 설치하려는 위치의 이미지에는 대상 구조물과 차량 등의 다른 구조물이 포함될 수 있어 구조물의 변위 응답 측정에 필요한 기준점을 인식하기 어려울 수 있다. 영상기반시스템의 설치 위치로 인해 차량이 포함된 왜곡된 이미지는 이미지 해석에서 부정확한 결과를 나타내므로 두 이미지 사이의 유사성을 산정하여 차량이 포함된 이미지를 제거하였다. 제안된 방법을 이용한 사장교 케이블의 장력을 측정하는 방법의 타당성을 확인하기 위해 공용 중인 사장교에서 진동법을 적용하여 장력을 측정하였다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년 학술대회 논문집 정보 및 제어부문
• /
• pp.527-529
• /
• 2006

In this paper, we find a robot's path using a Virtual Reality Modeling Language and overlay vision. For correct robot's path we describe a method for localizing a mobile robot in its working environment using a vision system and VRML. The robt identifies landmarks in the environment, using image processing and neural network pattern matching techniques, and then its performs self-positioning with a vision system based on a well-known localization algorithm. After the self-positioning procedure, the 2-D scene of the vision is overlaid with the VRML scene. This paper describes how to realize the self-positioning, and shows the overlap between the 2-D and VRML scenes. The method successfully defines a robot's path.

• International Journal of Advanced Culture Technology
• /
• 제6권3호
• /
• pp.201-210
• /
• 2018

We constructed 3-D images by HMD (Head Mount Display) type and polarization technique through which a training instrument of binocular function based on clinical vision therapy was developed and investigated its demand. To design the instrument that performs Test- Diagnosis - (Prescription) - Training - Treatment we provided a customized training method for the trainers using optical IT instruments of binocular test. Through the contents and the instruments of the vision therapy technology developed in this work we can replace the current vision test instruments to those costing 60% lower price in eye clinics and eye glasses shops. By replacing the existing vision test instruments it is possible to create a new market of the vision therapy in vision training instruments.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년 학술대회 논문집 정보 및 제어부문
• /
• pp.570-572
• /
• 2006

This paper discusses the method of moving object detection for biped robot walking. Most researches on vision based object detection have mostly focused on fixed camera based algorithm itself. However, developing vision systems for biped walking robot is an important and urgent issue since hired walking robots are ultimately developed not only for researches but to be utilized in real life. In the research, method for moving object detection has been developed for task assignment and execution of biped robot as well as for human robot interaction (HRI) system. But these methods are not suitable to biped walking robot. So, we suggest the advanced method which is suitable to biped walking robot platform. For carrying out certain tasks, an object detecting system using modified optical flow algorithm by wireless vision camera is implemented in a biped walking robot.