• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.2 s.308
• /
• pp.1-11
• /
• 2006

In this paper, we propose the novel video retrieval algorithm using non-parametric based motion classification in the shot-based video indexing structure. The proposed system firstly gets the key frame and motion information from each shot segmented by scene change detection method, and then extracts visual features and non-parametric based motion information from them. Finally, we construct real-time retrieval system supporting similarity comparison of these spatio-temporal features. After the normalized motion vector fields is created from MPEG compressed stream, the extraction of non-parametric based motion feature is effectively achieved by discretizing each normalized motion vectors into various angle bins, and considering a mean, a variance, and a direction of these bins. We use the edge-based spatial descriptor to extract the visual feature in key frames. Experimental evidence shows that our algorithm outperforms other video retrieval methods for image indexing and retrieval. To index the feature vectors, we use R*-tree structures.

• Journal of Broadcast Engineering
• /
• v.16 no.1
• /
• pp.108-122
• /
• 2011

In this paper, a transcoding method with low computational complexity and high coding efficiency is proposed to transcode distributed video coding (DVC) bitstreams to H.264/AVC ones. For the proposed high-performance transcoding with low complexity, not only Wyner-Ziv frames but also key frames can be transcoded with motion vectors estimated in generation of side information. As a motion vector is estimated from a key frame to a prior key frame for side information generation, the motion vector can be used to encode the intra key frame as a predicted frame. Motion estimation is performed with two predicted motion vectors. One is the motion vector from side information generation and the other is median of motion vectors of neighboring blocks. The proposed method selects the best motion vector between two motion vectors based on rate-distortion optimization. Coding efficiency can be improved with a small size of search range, because a motion vector estimated in side information generation is used as an initial motion vector for transcoding. In the experimental results, complexity of transcoder is reduced about 12% and bitrate performance increases about 28.7%.

• Proceedings of the IEEK Conference
• /
• 2004.06b
• /
• pp.585-588
• /
• 2004

This paper describes design and verification of the motion estimation and compensation unit using full search algorithm. Video processor is the key device of video communication systems. Motion estimation is the key module of video processor. The technologies of motion estimation and compensation unit are the core technologies for wireless video telecommunications system, portable multimedia systems. In this design, Verilog simulator and logic synthesis tools are used for hardware design and verification. In this paper, motion estimation and compensation unit are designed using FPGA, coded in Verilog HDL, and simulated and verified using Xilinx FPGA.

• Journal of KIISE:Software and Applications
• /
• v.27 no.5
• /
• pp.510-520
• /
• 2000

For content-based video indexing and retrieval, it is necessary to segment video data into video shots and then select key frames or representative frames for each shot. However, it is very difficult to select key frames automatically because the task of selecting meaningful frames is quite subjective. In this paper, we propose a new approach in selecting key frames based on visual contents such as region information and their temporal variations in the shot. First of all, we classify video shots into panning shots, zooming shots, tilting shots or no camera motion shots by detecting camera motion information in video shots. Then, in each category, we apply appropriate fuzzy rules to select key frames based on meaningful content in frame. Finally, we control the number of key frames in the selection process by adjusting the degree of detail in representing video shots.

• Earthquakes and Structures
• /
• v.17 no.6
• /
• pp.635-647
• /
• 2019

Tests and theoretical studies for seismic responses of a transmission tower-line system under coupled horizontal and tilt (CHT) ground motion were conducted. The method of obtaining the tilt component from seismic motion was based on comparisons from the Fourier spectrum of uncorrected seismic waves. The collected data were then applied in testing and theoretical analysis. Taking an actual transmission tower-line system as the prototype, shaking table tests of the scale model of a single transmission tower and towers-line systems under horizontal, tilt, and CHT ground motions were carried out. Dynamic equations under CHT ground motion were also derived. The additional P-∆ effect caused by tilt motion was considered as an equivalent horizontal lateral force, and it was added into the equations as the excitation. Test results were compared with the theoretical analysis and indicated some useful conclusions. First, the shaking table test results are consistent with the theoretical analysis from improved dynamic equations and proved its correctness. Second, the tilt component of ground motion has great influence on the seismic response of the transmission tower-line system, and the additional P-∆effect caused by the foundation tilt, not only increases the seismic response of the transmission tower-line system, but also leads to a remarkable asymmetric displacement effect. Third, for the tower-line system, transmission lines under ground motion weaken the horizontal displacement and acceleration responses of transmission towers. This weakening effect of transmission lines to the main structure, however, will be decreased with consideration of tilt component.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.7
• /
• pp.3599-3619
• /
• 2019

In human activity recognition system both static and motion information play crucial role for efficient and competitive results. Most of the existing methods are insufficient to extract video features and unable to investigate the level of contribution of both (Static and Motion) components. Our work highlights this problem and proposes Static-Motion fused features descriptor (SMFD), which intelligently leverages both static and motion features in the form of descriptor. First, static features are learned by two-stream 3D convolutional neural network. Second, trajectories are extracted by tracking key points and only those trajectories have been selected which are located in central region of the original video frame in order to to reduce irrelevant background trajectories as well computational complexity. Then, shape and motion descriptors are obtained along with key points by using SIFT flow. Next, cholesky transformation is introduced to fuse static and motion feature vectors to guarantee the equal contribution of all descriptors. Finally, Long Short-Term Memory (LSTM) network is utilized to discover long-term temporal dependencies and final prediction. To confirm the effectiveness of the proposed approach, extensive experiments have been conducted on three well-known datasets i.e. UCF101, HMDB51 and YouTube. Findings shows that the resulting recognition system is on par with state-of-the-art methods.

• ETRI Journal
• /
• v.37 no.2
• /
• pp.406-416
• /
• 2015

Timing plays a key role in expressing the qualitative aspects of a character's motion; that is, conveying emotional state, personality, and character role, all potentially without changing spatial positions. Temporal editing of locomotion style is particularly difficult for a novice animator since observers are not well attuned to the sense of weight and energy displayed through motion timing; and the interface for adjusting timing is far less intuitive to use than that for adjusting pose. In this paper, we propose an editing system that effectively captures the timing variations in an example locomotion set and utilizes them for style transfer from one motion to another via both global and upper-body timing transfers. The global timing transfer focuses on matching the input motion to the body speed of the selected example motion, while the upper-body timing transfer propagates the sense of movement flow - succession - through the torso and arms. Our transfer process is based on key times detected from the example set and transferring the relative changes of angle rotation in the upper body joints from a timing source to an input target motion. We demonstrate that our approach is practical in an interactive application such that a set of short locomotion cycles can be applied to generate a longer sequence with continuously varied timings.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.5
• /
• pp.998-1013
• /
• 2013

Wireless wearable sensor networks have emerged as a promising technique for human motion tracking due to the flexibility and scalability. In such system several wireless sensor nodes being attached to human limb construct a wearable sensor network, where each sensor node including MEMS sensors (such as 3-axis accelerometer, 3-axis magnetometer and 3-axis gyroscope) monitors the limb orientation and transmits these information to the base station for reconstruction via low-power wireless communication technique. Due to the energy constraint, the high fidelity requirement for real time rendering of human motion and tiny operating system embedded in each sensor node adds more challenges for the system implementation. In this paper, we discuss such challenges and experiences in detail during the implementation of such system with wireless wearable sensor network which includes COTS wireless sensor nodes (Imote 2) and uses TinyOS 1.x in each sensor node. Since our system uses the COTS sensor nodes and popular tiny operating system, it might be helpful for further exploration in such field.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.12 no.1
• /
• pp.376-386
• /
• 2020

The aim of this study was to develop a new efficient strategy that uses the Vector form Intrinsic Finite-element (VFIFE) method to conduct the static and dynamic analyses of marine pipes. Nonlinear problems, such as large displacement, small strain, and contact and collision, can be analyzed using a unified calculation process in the VFIFE method according to the fundamental theories of point value description, path element, and reverse motion. This method enables analysis without the need to integrate the stiffness matrix of the structure, because only motion equations of particles established according to Newton's second law are required. These characteristics of the VFIFE facilitate the modeling and computation efficiencies in analyzing the nonlinear dynamic problem of flexible pipe with large deflections. In this study, a three-dimensional (3-D) dynamical model based on 3-D beam element was established according to the VFIFE method. The deep-sea flexible pipe was described by a set of spatial mass particles linked by 3-D beam element. The motion and configuration of the pipe are determined by these spatial particles. Based on this model, a simulation procedure to predict the 3-D dynamical behavior of flexible pipe was developed and verified. It was found that the spatial configuration and static internal force of the mining pipe can be obtained by calculating the stationary state of pipe motion. Using this simulation procedure, an analysis was conducted on the static and dynamic behaviors of the flexible mining pipe based on a 1000-m sea trial system. The results of the analysis proved that the VFIFE method can be efficiently applied to the static and dynamic analyses of marine pipes.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.858-864
• /
• 2017

This study aims to restrain free conducting wire-type particles which are commonly and dangerously existing within DC gas-insulated transmission lines. A realistic platform of a coaxial cylindrical electrode was established by using a high-speed camera and a partial discharge (PD) monitor to observe the motion, PD, and breakdown of these particles. The probabilities of standing or bouncing, which can be affected by the length of the particles, were also quantitatively examined. The corona images of the particles were recorded, and particle-triggered PD signals were monitored and extracted. Breakdown images were also obtained. The air-gap breakdown with the particles was subjected to mechanism analysis on the basis of stream theory. Results reveal that the lifting voltage of the wire particles is almost irrelevant to their length but is proportional to the square root of their radius. Short particles correspond to high bouncing probability. The intensity and frequency of PD and the micro-discharge gap increase as the length of the particles increases. The breakdown voltage decreases as the length of the particles decreases.