• Journal of Satellite, Information and Communications
• /
• v.9 no.4
• /
• pp.32-37
• /
• 2014

The panorama is a good alternative to overcome narrow FOV under study in robot vision, stereo camera and panorama image registration and modeling. The panorama can materialize view with angles wider than human view and provide realistic space which make feeling of being on the scene based on realism. If we use all correspondence, it is too difficult to find strong features and correspondences and assume accurate homography matrix in geographic changes in images as load of calculation increases. Accordingly, we used SURF algorithm to estimate overlapping areas with high similarity by comparing and analyzing the input images' histograms and to detect features. And we solved the problem of input order so we can make panorama by input images without order.

• Journal of Information Management
• /
• v.24 no.4
• /
• pp.58-77
• /
• 1993

Database of journals/books are categorized into bibliographic database and clearinghouse-type databases which tell you locations of original materials. There is such a problem in the latter type of databases that those are not likely to be commercialized, although ultimate purpose which users have is to obtain original materials. We find photocopying service through on-line or full-text databases currently available, but we can't get information which meets our needs only by those databases. This paper describes major database services and how to use them by document type(journal or book). The author also discusses what future information centers utilizing databases should be.

• Journal of IKEEE
• /
• v.24 no.4
• /
• pp.1109-1116
• /
• 2020

The matching score evaluating human copying ability can be a good measure to check children's developmental stages, or sports movements like golf swing and dance, etc. It also can be used as HCI for AR, VR applications. This paper presents a method to evaluate the motion similarity between demonstrator who initiates movement and participant who follows the demonstrator action. We present a quantification method of the similarity which utilizes Euclidean L2 distance of Openpose keypoins vector similarity. The proposed method adapts DTW, thus can flexibly cope with the time delayed motions.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.5
• /
• pp.61-68
• /
• 2021

In this paper, we propose an edge computing-based exercise data collection device that can be provided for exercise rehabilitation services. In the existing cloud computing method, when the number of users increases, the throughput of the data center increases, causing a lot of delay. In this paper, we design and implement a device that measures and estimates the position of keypoints of body joints for movement information collected by a 3D camera from the user's side using edge computing and transmits them to the server. This can build a seamless information collection environment without load on the cloud system. The results of this study can be utilized in a personalized rehabilitation exercise coaching system through IoT and edge computing technologies for various users who want exercise rehabilitation.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.97-99
• /
• 2021

본 논문에서는 딥러닝 기반 특징점 필터링 방법을 이용한 원격 탐사 영상에 대한 영상 정합 (Image Registration) 고속화 방법을 제안한다. 기존의 특징 기반 영상 정합 방법의 복잡도는 특징 매칭 (Feature Matching) 단계에서 발생한다. 이 복잡도를 줄이기 위하여 본 논문에서는 특징 매칭이 영상의 인공구조물에서 검출된 특징점으로 매칭되는 것을 확인하여 특징점 검출기에서 검출된 특징점 중에서 인공구조물에서 검출된 특징점만 필터링하는 방법을 제안한다. 딥러닝 기반 특징점 필터링은 영상 정합을 위하여 필수적인 특징점을 잃지 않으면서 그 수를 줄이기 위하여 인공구조물의 경계와 인접한 특징점을 보존하고, 축소한 영상을 사용하며, 영상 분할(Image Segmentation) 방법의 결과에서 생기는 영상 패치 경계의 잡음을 제거하기 위하여 영상 패치를 중복하여 잘라 냄으로써 정합 속도와 정확도를 향상시킨다. 영상 정합 고속화 방법을 의 성능을 검증하기 위하여 아리랑 3 호 위성 원격 탐사 영상을 사용하여 기존 특징점 추출 방법과 속도와 정확도를 비교하였다. 딥러닝 기반 영상 정합 방법을 기준으로 하여 비교하였을 때 특징점의 수를 약 82% 감소시키면서 속도를 약 9.17 배 향상시켰지만 정확도가 0.985 에서 0.855 으로 저하되었다.

• International Journal of Internet, Broadcasting and Communication
• /
• v.15 no.4
• /
• pp.142-148
• /
• 2023

In Modern days, Self-driving for modern people is an absolute necessity for transportation and many other reasons. Additionally, after the outbreak of COVID-19, driving by oneself is preferred over other means of transportation for the prevention of infection. However, due to the constant exposure to stressful situations and chronic fatigue one experiences from the work or the traffic to and from it, modern drivers often drive under drowsiness which can lead to serious accidents and fatality. To address this problem, we propose a drowsy driving prevention learning model which detects a driver's state of drowsiness. Furthermore, a method to sound a warning message after drowsiness detection is also presented. This is to use MoveNet to quickly and accurately extract the keypoints of the body of the driver and Dense Neural Network(DNN) to train on real-time driving behaviors, which then immediately warns if an abnormal drowsy posture is detected. With this method, we expect reduction in traffic accident and enhancement in overall traffic safety.

• KIPS Transactions on Computer and Communication Systems
• /
• v.3 no.5
• /
• pp.163-172
• /
• 2014

The Scale Invariant Feature Transform (SIFT) has been widely used in a lot of applications for image feature matching. Such a transform allows us to strong matching ability, stability in rotation, and scaling with the variety of different scales. Recently, it has been made one of the most successful algorithms in the research areas of copy-move forgery detections. Though this transform is capable of identifying copy-move forgery, it does not widely address the possibility that counter-forensics operations may be designed and used to hide the evidence of image tampering. In this paper, we propose a targeted counter-forensics method for impeding SIFT-based copy-move forgery detection by applying a semantically admissible distortion in the processing tool. The proposed method allows the attacker to delude a similarity matching process and conceal the traces left by a modification of SIFT keypoints, while maintaining a high fidelity between the processed images and original ones under the semantic constraints. The efficiency of the proposed method is supported by several experiments on the test images with various parameter settings.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.46 no.3
• /
• pp.86-95
• /
• 2009

SIFT is an algorithm to extract vectors at pixels around keypoints, in which the pixel colors are very different from neighbors, such as vertices and edges of an object. The SIFT algorithm is being actively researched for various image processing applications including 3D image reconstructions and intelligent vision system for robots. In this paper, we implement a hardware to sift feature detection algorithm for real time processing in embedded systems. We estimate that the hardware implementation give a performance 25ms of $1,280{\times}960$ image and 5ms of $640{\times}480$ image at 100MHz. And the implemented hardware consumes 45,792 LUTs(85%) with Synplify 8.li synthesis tool.

• Journal of KIISE:Computing Practices and Letters
• /
• v.14 no.5
• /
• pp.477-481
• /
• 2008

Place recognition is necessary for a mobile user to be provided with place-dependent information. This paper proposes real-time video based place recognition system that identifies users' current place while moving in the building. As for the feature extraction of a scene, there have been existing methods based on global feature analysis that has drawback of sensitive-ness for the case of partial occlusion and noises. There have also been local feature based methods that usually attempted object recognition which seemed hard to be applied in real-time system because of high computational cost. On the other hand, researches using statistical methods such as HMM(hidden Markov models) or bayesian networks have been used to derive place recognition result from the feature data. The former is, however, not practical because it requires huge amounts of efforts to gather the training data while the latter usually depends on object recognition only. This paper proposes a combined approach of global and local feature analysis for feature extraction to complement both approaches' drawbacks. The proposed method is applied to a mobile information system and shows real-time performance with competitive recognition result.

• Journal of Information Management
• /
• v.25 no.1
• /
• pp.63-79
• /
• 1994

Analysis systems and databases covering economics and industrial areas in Japan have long history, but due to considerable depth in specialty those have been used only by limited, certain patrons. It means that there are still few of practical system services at full scale. This paper describes the representative system services. Taking an example of interest prediction the author explains now to utilize systems and some points to be reminded. He also describes how to confirm newspaper information, how to predict economics, how to use various kinds of models based on economic prediction, and industrial analysis. Researches and studies are very often proceeded on economic prediction, and industrial analysis. Researches and studies are very often proceeded through interaction among researchers. So that we should make efforts continuously such as to rountinely get familiar with systems, to exchange information among users, to utilize helpdesks every time we need.