• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.5
• /
• pp.55-64
• /
• 2018

In this paper, we propose fully convolutional neural networks based automatic detection of a pothole that directly causes driver's safety accidents and the vehicle damage. First, the training DB is collected through the camera installed in the vehicle while driving on the road, and the model is trained in the form of a semantic segmentation using the fully convolutional neural networks. In order to generate robust performance in a dark environment, we augmented the training DB according to brightness, and finally generated a total of 30,000 training images. In addition, a total of 450 evaluation DB was created to verify the performance of the proposed automatic pothole detection, and a total of four experts evaluated each image. As a result, the proposed pothole detection showed robust performance for missing.

• Journal of the Korea Society of Computer and Information
• /
• v.11 no.6 s.44
• /
• pp.133-141
• /
• 2006

Recently many wearable computers have been developed. But they still have many user interface problems from both an input and output perspective. This paper presents a wearable user interface based on EOG(electrooculogram) sensing circuit and marker recognition. In the proposed user interface, the EOG sensor circuit which tracks the movement of eyes by sensing the potential difference across the eye is used as a pointing device. Objects to manipulate are represented human readable markers. And the marker recognition system detects and recognize markers from the camera input image. When a marker is recognized, the corresponding property window and method window are displayed to the head mounted display. Users manipulate the object by selecting a property or a method item from the window. By using the EOG sensor circuit and the marker recognition system, we can manipulate an object with only eye movement in the wearable computing environment.

• Journal of Hydrogen and New Energy
• /
• v.26 no.3
• /
• pp.287-293
• /
• 2015

The current study has investigated the effects of biodiesel blended with gasoline on the spray characteristics in a Constant Volume Combustion Chamber (CVCC). With the concentration of 5, 10, 15 and 20% by volume, biodiesel was blended with commercial gasoline and performed on the macroscopic visualization test. Pure gasoline and biodiesel were also tested as the reference. The shadowgraph technique was conducted in the constant volume chamber. The spray images were recorded by a high speed video camera with frame speed 10,000 frame per second. Fuel injection was set at 800, 1000 and 1,350 bar with the simulated speed 1,500 and 2,000 rpm. The back pressure was controlled at 20 bar. The spray angle and penetration tip were measured and analyzed by using the image processing. At the high injection pressure, the spray penetration length with the simulated speed 1,500 rpm showed that B100 was lower than GB00-20 whereas the spray penetration length with the simulated speed 2,000 rpm exhibited that GB blends and B100 were insignificantly different. Due to biodiesel concentration, its effects on spray angles were observed throughout injection periods (T1, T2 and T3). At the simulated speed 1,500 rpm, the spray angle of GB blends and B100 presented the same pattern following injection timing. In addition, when the simulated speed increased to 2,000 rpm the different spray angle of all blends disappeared at main injection (T3).

• Physical Therapy Korea
• /
• v.24 no.4
• /
• pp.46-53
• /
• 2017

Background: The MLA is supported by both the abductor hallucis (ABH) and the extrinsic muscles. Insufficient muscular support may lower the MLA when the body's weight is applied to the foot. The short foot exercise (SFE) is effective in increasing the height of the MLA for people with flat feet. Most of the research related to the SFE has simply evaluated the efficiency of the exercise using enhanced ABH electromyography (EMG) activation. Since the tibialis anterior (TA), peroneus longus (PER), and ABH are all involved in supporting the MLA, a new experiment design examining the EMG of the selected muscles during SFE should be applied to clarify its effect. Objects: Therefore, this study aimed to clarify the effect of the SFE in different ankle position on the MLA angle and the activation of both the intrinsic and extrinsic muscles and to determine the optimal position. Methods: 20 healthy subjects and 12 subjects with flat feet were recruited from Yonsei University. The surface EMG and camera were used to collect muscle activation amplitude of TA, PER, and ABH and to capture the image of MLA angle during SFE. The subjects performed the SFE while sitting in three different ankle positions-neutral (N), dorsiflexion (DF) at 30 degrees, and plantar flexion (PF) at 30 degrees. Results: ABH EMG amplitudes were significantly greater in N and DF than in PF (p<.05). Muscle activation ratio of TA to ABH was the lowest in PF (p<.05). MLA angle in both groups significantly decreased in PF (p<.01). The TA and ABH was activated at the highest level in DF. However, in PF, subjects significantly activated the ABH and PER with relatively low activation of TA. Conclusion: Therefore, researchers need to discuss which SFE condition most effectively use the arch support muscle for flat foot.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.8B
• /
• pp.1577-1585
• /
• 1999

In this paper, we propose detection methods for gradual scene changes such as dissolve, pan, and zoom. The proposal method to detect a dissolve region uses scene features based on spatial statistics of the image. The spatial statistics to define shot boundaries are derived from squared means within each local area. We also propose a method of the camera motion detection using four representative motion vectors in the background. Representative motion vectors are derived from macroblock motion vectors which are directly extracted from MPEG streams. To reduce the implementation time, we use DC sequences rather than fully decoded MPEG video. In addition, to detect the gradual scene change region precisely, we use all types of the MPEG frames(I, P, B frame). Simulation results show that the proposed detection methods perform better than existing methods.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.10 no.4
• /
• pp.91-97
• /
• 2010

This paper describes a development of an network based remote surveillance system using omni-directional mobile robot. the proposed surveillance system can control a mobile robot to move and examines the given place closely while the conventional surveillance system uses a fixed camera. The mobile robot in the proposed system has three omni-directional wheels to move to any given direction freely. We also developed the proposed system as robot services using Microsoft's MSRDS for a user to control the mobile robot and monitor the remote scene captured from the robot. Finally we verified the feasibility and effectiveness of the proposed system by conducting the remote operating the mobile robot and monitoring experiments in a networked environment. We also conducted a color based object detection and motion detection on image sequences acquired from a remote mobile robot in an another PC in a network environment.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.10a
• /
• pp.553-559
• /
• 2004

In todays security industry, personal identification is also based on biometric. Biometric identification is performed basing on the measurement and comparison of physiological and behavioral characteristics, Biometric for recognition includes voice dynamics, signature dynamics, hand geometry, fingerprint, iris, etc. Iris can serve as a kind of living passport or living password. Iris recognition system is the one of the most reliable biometrics recognition system. This is applied to client/server system such as the electronic commerce and electronic banking from stand-alone system or networks, ATMs, etc. A new algorithm using nonlinear function in recognition process is proposed in this paper. An algorithm is proposed to determine the localized iris from the iris image received from iris input camera in client. For the first step, the algorithm determines the center of pupil. For the second step, the algorithm determines the outer boundary of the iris and the pupillary boundary. The localized iris area is transform into polar coordinates. After performing three times Wavelet transformation, normalization was done using sigmoid function. The converting binary process performs normalized value of pixel from 0 to 255 to be binary value, and then the converting binary process is compare pairs of two adjacent pixels. The binary code of the iris is transmitted to the by server. the network. In the server, the comparing process compares the binary value of presented iris to the reference value in the University database. Process of recognition or rejection is dependent on the value of Hamming Distance. After matching the binary value of presented iris with the database stored in the server, the result is transmitted to the client.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.2
• /
• pp.184-189
• /
• 2004

In this paper, we realized a system that receives the dynamic images of finger languages, which is the method of intention transmission of the hearing disabled person, using the white and black CCD camera, and that recognizes the images and converts them to the editable text document. We use the afterimage to draw a sharp line between indistinct images and clear images from a series of inputted images, and get the character alphabet from the away of continuous images and output the accomplished character to the word editor by applying the automata theory. After the system removes the varied wrist part from the data of clean image, it gets the controid point of hand by the maximum circular movement method and recognizes the hand that is necessary to analyze the finger languages by applying the circular pattern vector algorithm. The system abstracts the characteristic vectors of the hand using the distance spectrum from the center of the hand and it compares the characteristic vector of inputted pattern from the standard pattern by applying the fuzzy inference and recognizes the movement of finger languages.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.7
• /
• pp.838-846
• /
• 2004

In the case where both a camera and a target are moving at the same time, the image background is successively changed, and the overlap with other moving objects is apt to be generated. The snake algorithms have been variously used in tracking the object, but it is difficult to be applied in the excessive overlap with other objects and the large bias between the snake and the target. To solve this problem, this paper presents an extended snake model. It includes an additional energy function which considers the temporal variation rate of the snake's area and a SSD algorithm which generates the template adaptive to the snake detected in the previous frame. The new energy function prevents the snake from over-shrinking or stretching and the SSD algorithm with adaptively changing template allows the prediction of the target's position in the next frame. The experimental results have shown that the proposed algorithm successfully tracks the target even when the target is temporarily occluded by other objects.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.8
• /
• pp.1693-1700
• /
• 2012

Recently, the warning system to aid drivers for safe driving is being developed. The system estimates the distance between the driver's car and the car before it and informs him of safety distance. In this paper, we designed and implemented the collision warning system which detects the car in front on the actual road situation and measures the distance between the cars in order to detect the risk situation for collision and inform the driver of the risk of collision. First of all, using the forward-looking camera, it extracts the interest area corresponding to the road and the cars from the image photographed from the road. From the interest area, it extracts the object of the car in front through the analysis on the critical value of the shadow of the car in front and then alerts the driver about the risk of collision by calculating the distance from the car in front. Based on the results of detecting driving cars and measuring the distance between cars, the collision warning system was designed and realized. According to the result of applying it in the actual road situation and testing it, it showed very high accuracy; thus, it has been verified that it can cope with safe driving.