• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.3C
• /
• pp.125-132
• /
• 2005

This paper describes a computer vision system based on active IR illumination for real-time gaze discrimination system. Unlike most of the existing gaze discrimination techniques, which often require assuming a static head to work well and require a cumbersome calibration process for each person, our gaze discrimination system can perform robust and accurate gaze estimation without calibration and under rather significant head movement. This is made possible by a new gaze calibration procedure that identifies the mapping from pupil parameters to screen coordinates using generalized regression neural networks (GRNNs). With GRNNs, the mapping does not have to be an analytical function and head movement is explicitly accounted for by the gaze mapping function. Futhermore, the mapping function can generalize to other individuals not used in the training. To further improve the gaze estimation accuracy, we employ a reclassification scheme that deals with the classes that tend to be misclassified. This leads to a 10% improvement in classification error. The angular gaze accuracy is about 5°horizontally and 8°vertically. The effectiveness of our gaze tracker is demonstrated by experiments that involve gaze-contingent interactive graphic display.

• Journal of Electrical Engineering and Technology
• /
• v.10 no.6
• /
• pp.2221-2227
• /
• 2015

To arrest a frequency nadir, a stepwise inertial control (SIC) scheme generates a constant active power reference signal of a wind turbine generator (WTG) immediately after a disturbance and maintains it for the predetermined time. From that point, however, the reference of a WTG abruptly decreases to restore the rotor speed for the predefined period. The abrupt decrease of WTG output power will inevitably cause a second frequency dip. In this paper, we propose a modified SIC scheme of a doubly-fed induction generator (DFIG) that can prevent a second frequency dip. A reference value of the modified SIC scheme consists of a reference for the maximum power point tracking control and a constant value. The former is set to be proportional to the cube of the rotor speed; the latter is determined so that the rotor speed does not reach the minimum operating limit by considering the mechanical power curve of a DFIG. The performance of the modified SIC was investigated for a 100 MW aggregated DFIG-based wind power plant under various wind conditions using an EMTP-RV simulator. The results show that the proposed SIC scheme significantly increases the frequency nadir without causing a second frequency dip.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.3 no.1
• /
• pp.38-46
• /
• 2010

This paper describes a computer vision system based on active IR illumination for real-time gaze discrimination system. Unlike most of the existing gaze discrimination techniques, which often require assuming a static head to work well and require a cumbersome calibration process for each person, our gaze discrimination system can perform robust and accurate gaze estimation without calibration and under rather significant head movement. This is made possible by a new gaze calibration procedure that identifies the mapping from pupil parameters to screen coordinates using generalized regression neural networks (GRNNs). With GRNNs, the mapping does not have to be an analytical function and head movement is explicitly accounted for by the gaze mapping function. Furthermore, the mapping function can generalize to other individuals not used in the training. To further improve the gaze estimation accuracy, we employ a reclassification scheme that deals with the classes that tend to be misclassified. This leads to a 10% improvement in classification error. The angular gaze accuracy is about $5^{\circ}$horizontally and $8^{\circ}$vertically. The effectiveness of our gaze tracker is demonstrated by experiments that involve gaze-contingent interactive graphic display.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.2
• /
• pp.322-329
• /
• 2005

This paper describes a computer vision system based on active IR illumination for real-time gaze discrimination system. Unlike most of the existing gaze discrimination techniques, which often require assuming a static head to work well and require a cumbersome calibration process for each person, our gaze discrimination system can perform robust and accurate gaze estimation without calibration and under rather significant head movement. This is made possible by a new gaze calibration procedure that identifies the mapping from pupil parameters to screen coordinates using generalized regression neural networks (GRNNS). With GRNNS, the mapping does not have to be an analytical function and head movement is explicitly accounted for by the gaze mapping function. furthermore, the mapping function can generalize to other individuals not used in the training. To further improve the gaze estimation accuracy, we employ a reclassification scheme that deals with the classes that tend to be misclassified. This leads to a 10$\%$ improvement in classification error. The angular gaze accuracy is about $5^{circ}$horizontally and $8^{circ}$vertically. The effectiveness of our gaze tracker is demonstrated by experiments that involve gaze-contingent interactive graphic display.

• The Journal of the Korea Contents Association
• /
• v.18 no.3
• /
• pp.11-18
• /
• 2018

To interact with a virtual object in immersive virtual environment, the target object should be selected quickly and accurately. Conventional 3D ray casting method using a direction of user's hand or head allows the user to select an object quickly. However, accuracy problem occurs when selecting an object using conventional methods among occlusion of objects. In this paper, we propose a region of interest based selection method that enables to select an object among occlusion of objects using a combination of gaze tracking and hand gesture recognition. When a user looks at a group of occlusion of objects, the proposed method recognizes user's gaze input, and then region of interest is set by gaze input. If the user wants to select an object among them, the user gives an activation hand gesture. Then, the proposed system relocates and visualizes all objects on a virtual active window. The user can select an object by a selecting hand gesture. Our experiment verified that the user can select an object correctly and accurately.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1202-1205
• /
• 2005

This paper describes a method for vision-based person identification that can detect, track, and recognize person from video using multiple cues: height and dressing colors. The method does not require constrained target's pose or fully frontal face image to identify the person. First, the system, which is connected to a pan-tilt-zoom camera, detects target using motion detection and human cardboard model. The system keeps tracking the moving target while it is trying to identify whether it is a human and identify who it is among the registered persons in the database. To segment the moving target from the background scene, we employ a version of background subtraction technique and some spatial filtering. Once the target is segmented, we then align the target with the generic human cardboard model to verify whether the detected target is a human. If the target is identified as a human, the card board model is also used to segment the body parts to obtain some salient features such as head, torso, and legs. The whole body silhouette is also analyzed to obtain the target's shape information such as height and slimness. We then use these multiple cues (at present, we uses shirt color, trousers color, and body height) to recognize the target using a supervised self-organization process. We preliminary tested the system on a set of 5 subjects with multiple clothes. The recognition rate is 100% if the person is wearing the clothes that were learned before. In case a person wears new dresses the system fail to identify. This means height is not enough to classify persons. We plan to extend the work by adding more cues such as skin color, and face recognition by utilizing the zoom capability of the camera to obtain high resolution view of face; then, evaluate the system with more subjects.

• The Journal of the Korea Contents Association
• /
• v.11 no.1
• /
• pp.92-101
• /
• 2011

However, in order to create a stereoscopic 3D image, independently, it is required to spend expensive manufacturing costs and to have special techniques. As 3D display devices have been generalized, there is an increasing need for implementing a stereoscopic 3D image without a burden of expensive costs. This paper proposes some methods to implement stereoscopic 3D images easily by utilizing a marker tracking technology using a single camera. In addition, the study made it possible for the resolution of an image to be adjustable dynamically. This paper will be committed to the promotion of the field of UCC (User Created Contents) using stereoscopic 3D images by attracting the active participation of general users in the field of the implementation of stereoscopic 3D images.

• Journal of the HCI Society of Korea
• /
• v.5 no.2
• /
• pp.1-6
• /
• 2010

In this research, we designed the head mouse system for disabled and gamers, a mouse controller which can be controlled by head movements and eye blinks only, and compared its performance with other regular mouse controller systems. The head mouse was moved by a gyro-sensor, which can measure an angular rotation of a head movement, and the eye blink was used as a clicking event of the mouse system. Accumulated errors caused by integral, which was a problem that previous head mouse system had, were removed periodically, and treated as dead zones in the non-linear relative point graph, and direct mouse point control was possible using its moving distance and acceleration calculation. We used the active light sources to minimize the influence of the ambient light changes, so that the head mouse was not affected by the change in external light source. In a comparison between the head mouse and the gazing tracking mouse (Quick Glance), the above method resulted about 21% higher score on the clicking event experiment called "20 clicks", about 25% higher on the dasher experiment, and about 37% higher on on-screen keyboard test respectively, which concludes that the proposed head mouse has better performance than the standard mouse system.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.6
• /
• pp.17-25
• /
• 2001

Traditional adaptive noise cancelling methods rely their performance on various interfering parameters, such as convergence speed, tracking ability, numerical stability, relative frequency characteristics between target and reverberation signals, and activity of the target. In this paper, an adaptive noise cancelling method is suggested, which Provides a successful tradeoff mon these factors. It is designed to work on the transform domain, adopts the Gram-Schmidt orthogonalization process, and is implemented by the escalator algorithm. The transform domain approach supports a tradeoff between the convergence speed and numerical cost. The proposed method is verified by applying a real-data collected in the shallow waters off the east coasts of korea. It is shown that it has a good reverberation-rejection capability even for the target signal with adjacent frequency components to those of the reverberation, and its performance is invariant for the activity of the target.

• Journal of Environmental Science International
• /
• v.29 no.9
• /
• pp.885-905
• /
• 2020

This study described the characteristics of the upstream migration of salmon (Chum salmon, Oncorhynchus keta) along Yeon-gok Stream in the eastern coastal region of Korea from October 24 to November 9, 2018 using radio tag and data storage tag loggers for the detection of the locations of tagged salmon and measurement of water temperature. Tracking experiments were conducted and classified into four types (case 1 to case 4) depending on the release time and the number of salmon tracked. Experiments from case 1 to case 3 were classified depending on the number of salmon tracked into cases in which a single tagged salmon was tracked (case 1), a pair of tagged salmon was tracked (case 2), and salmon were tracked by different sex ratios (case 3). Experiments from cases 1 to 3 were conducted between 10 AM and 1 PM, and case 4 was conducted after 3:30 PM. Salmon moved and spawned in the downstream region of the Yeon-gok, where water temperature is higher than in other rivers and salmon return in Canada, Russia, Japan, and the U.S.A. Most of the radio-tagged salmon swam in deep and shaded areas during the day but actively moved upstream close to sunset, regardless of the release time. Females showed relatively more active movements than males during upstream migration.