• Proceedings of the KIEE Conference
• /
• 2007.04a
• /
• pp.301-303
• /
• 2007

This paper presents user adaptation methods to overcome limitations of a user-independent model and a user-dependent model in a Korean sign language recognition system. To adapt model parameters for unobserved states in hidden Markov models, we introduce new methods based on motion similarity and prediction from adaptation history so that we can achieve faster adaption and higher recognition rates comparing with previous methods.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.5 no.4
• /
• pp.257-265
• /
• 1981

A modified subsonic wind tunnel was employed to investigate the turbulent flow of a valve jet. To effectuate this experimental study, an actual valve and valve seat of a diesel engine were equipped at its outlet (ref. Fig. 3) Theoretically, using the equation of motion of Navier-Stokes in the chlindrical coordinates, the turbulent equation of motion for the incompressible fluid was derived with three assumptions; steady flow (.delta./.delta.t=0), axisymmetry and revolutionary homogeneity(.delta./.delta..phi.=0), no ensemble revolution (V.iden.0). Experimentally, mean and fluctuation velocities have been measured in the redial direction. With a assumption of a non-dimensional velocity distribution equation, a sami-emperical similarity profile equation of the jet flow have been derived, whose empirical constants were deterimed graphically with the data obtained.

• IEIE Transactions on Smart Processing and Computing
• /
• v.4 no.6
• /
• pp.434-442
• /
• 2015

In this paper, we propose an integrated spatial and view scalable video codec based on high efficiency video coding (HEVC). The proposed video codec is developed based on similarity and uniqueness between the scalable extension and 3D multi-view extension of HEVC. To improve compression efficiency using the proposed scalable multi-view video codec, inter-layer and inter-view predictions are jointly employed by using high-level syntaxes that are defined to identify view and layer information. For the inter-view and inter-layer predictions, a decoded picture buffer (DPB) management algorithm is also proposed. The inter-view and inter-layer motion predictions are integrated into a consolidated prediction by harmonizing with the temporal motion prediction of HEVC. We found that the proposed scalable multi-view codec achieves bitrate reduction of 36.1%, 31.6% and 15.8% on the top of ${\times}2$, ${\times}1.5$ parallel scalable codec and parallel multi-view codec, respectively.

• Journal of Ocean Engineering and Technology
• /
• v.13 no.2 s.32
• /
• pp.129-137
• /
• 1999

The time of the onset of double-diffusive convection in time-dependent, nonlinear concentration fields is investigated theoretically. The initially quiescent horizontal fluid layer with a uniform temperature gradient experiences a sudden concentration change from below, but its stable thermal stratification affects concentration effects in such way to invoke convective motion. The related stability analysis, including Soret effect, is conducted on the basis of the propagation theory. Under the linear stability theory the concentration penetration depth is used as a length scaling factor, and the similarity transform for the linearized perturbation equations. The newlly obtained stability equations are solved numerically. The resulting critical time to mark the onset of regular cells are obtained as a function of the thermal Rayleigh number, the solute Rayleigh number, and the Soret effect coefficient. For a certain value of the Soret effect coefficient, the stable thermal gradient promote double-diffusive convective motion.

• Journal of Information Processing Systems
• /
• v.19 no.1
• /
• pp.80-88
• /
• 2023

RGB-thermal (RGB-T) tracking using unmanned aerial vehicles (UAVs) involves challenges with regards to the similarity of objects, occlusion, fast motion, and motion blur, among other issues. In this study, we propose dynamic tracking aggregation (DTA) as a unified framework to perform object detection and data association. The proposed approach obtains fused features based a transformer model and an L1-norm strategy. To link the current frame with recent information, a dynamically updated embedding called dynamic tracking identification (DTID) is used to model the iterative tracking process. For object association, we designed a long short-term tracking aggregation module for dynamic feature propagation to match spatial and temporal embeddings. DTA achieved a highly competitive performance in an experimental evaluation on public benchmark datasets.

• Journal of KIISE:Databases
• /
• v.35 no.4
• /
• pp.345-355
• /
• 2008

SIFT is popularly used in computer vision application such as object recognition, motion tracking, and 3D reconstruction among various shape descriptors. However, it is not easy to apply SIFT into the image similarity search as it is since it uses many high dimensional keypoint vectors. In this paper, we present a SIFT based image similarity search method using an edge image pyramid and an interesting region detection. The proposed method extracts keypoints, which is invariant to contrast, scale, and rotation of image, by using the edge image pyramid and removes many unnecessary keypoints from the image by using the hough transform. The proposed hough transform can detect objects of ellipse type so that it can be used to find interesting regions. Experimental results demonstrate that the retrieval performance of the proposed method is about 20% better than that of traditional SIFT in average recall.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.5
• /
• pp.40-47
• /
• 2009

A motion adaptive video noise reduction scheme is proposed by cascading a temporal filter and a spatial filter. After a noise-robust motion detection is performed with a pre-spatial filter, the strength of the motion adaptive temporal filter is controlled by the amount of temporal movement. In order to fully utilize the temporal correlation of video signal, noisy input image is processed first by the temporal filter, therefore, image details of temporally stationary region are quite well preserved while undesired noises are suppressed. In contrast to the pre-spatial filter used for the robust motion detection, the cascaded post-spatial filter removes the remained noises by considering the strength of the temporal filter and the spatial self-similarity search results obtained from the pre-spatial filter.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.8
• /
• pp.2866-2880
• /
• 2014

We propose a video denoising method based on Kalman filter to reduce the noise in video sequences. Firstly, with the strong spatiotemporal correlations of neighboring frames, motion estimation is performed on video frames consisting of previous denoised frames and current noisy frame based on intensity and structure tensor. The current noisy frame is processed in temporal domain by using motion estimation result as the parameter in the Kalman filter, while it is also processed in spatial domain using the Wiener filter. Finally, by weighting the denoised frames from the Kalman and the Wiener filtering, a satisfactory result can be obtained. Experimental results show that the performance of our proposed method is competitive when compared with state-of-the-art video denoising algorithms based on both peak signal-to-noise-ratio and structural similarity evaluations.

• Journal of Advanced Navigation Technology
• /
• v.10 no.4
• /
• pp.400-405
• /
• 2006

In this paper, we propose a moving object tracking algorithm by using motion estimation in UAV(Unmanned Aerial Vehicle) video. Proposed algorithm is based on generation of initial image from detected reference image, and tracking of moving object under the time-varying image. With a series of this procedure, tracking process is stable even when the UAV camera sways by correcting position of moving object, and tracking time is relatively reduced. A block matching algorithm is also utilized to determine the similarity between reference image and moving object. An experimental result shows that our proposed algorithm is better than the existing full search algorithm.

• The Academic Congress of Korean Shoulder and Elbow Society
• /
• 2009.03a
• /
• pp.175-175
• /
• 2009

Despite its importance for the understanding of joint kinematics in vivo, there has been few studies about shoulder joints. The purpose of this study is to analyze the glenohumeral joint during internal and external rotation at 90 degrees of abduction using in vivo noninvasive motion analysis system. MRI was performed for the following seven positions from maximum internal rotation to maximum external rotation at intervals of 30 degrees. We used 3D-gradient echo sequencing (TR: 12 ms, TE: 5.8 ms, 0.8 mm-slice thickness). Our method is based on matching three-dimensional MR images by the similarity of the image intensity. We analyzed the in vivo three-dimensional motions of the glenohumeral and scapulothoracic joint during this motion. In scapla plane, the mean rotation angle of the glenohumeral join was 105.5 degrees ($SD{\pm}39.0^{\circ}$). The mean rotation angle of the scapulothracic joint was 27.5 degrees ($SD\;{\pm}\;7.7^{\circ}$). The contribution ratio is almost 3.8:1 of glenohumeral and scapulothracic joint respectively.