Human action recognition is an active topic in computer vision. In recent years, tangible results... more Human action recognition is an active topic in computer vision. In recent years, tangible results have been shown through deep learning methods; however, at a very high computational cost. They may be suitable for video retrieval or video summarization of a movie or drama, but they are not suitable for the visual surveillance of human action, which should be analyzed in real time. In this study, we propose an action change detection method to reduce the computational cost. On the surveillance camera screen, anomalous actions such as movies and sports are frequently not observed, and simple actions are often repeated. Thus, it is very inefficient to continue to apply high cost action recognition methods on repeated actions. The proposed action change detection method decides whether the previous action of the person is the same as the current action. The action recognition method is applied only when it has determined that the action has changed. The action change detection process i...
Abstract In pattern recognition, the classification accuracy has a strong correlation with the se... more Abstract In pattern recognition, the classification accuracy has a strong correlation with the selected features. Therefore, in the present paper, we applied an evolutionary algorithm in combination with linear discriminant analysis (LDA) to enhance the feature selection in a static image-based facial expressions system. The accuracy of the classification depends on whether the features are well representing the expression or not. Therefore the optimization of the selected features will automatically improve the classification accuracy. The proposed method not only improves the classification but also reduces the dimensionality of features. Our approach outperforms linear-based dimensionality reduction algorithms and other existing genetic-based feature selection algorithms. Further, we compare our approach with VGG (Visual Geometry Group)-face convolutional neural network (CNN), according to the experimental results, the overall accuracy is 98.67% for either our approach or VGG-face. However, the proposed method outperforms CNN in terms of training time and features size. The proposed method proves that it is able to achieve high accuracy by using far fewer features than CNN and within a reasonable training time.
Behavior of intelligent video surveillance system is recognized by analyzing the pattern of the o... more Behavior of intelligent video surveillance system is recognized by analyzing the pattern of the object of interest by using the frame information of video inputted from the camera and analyzes the behavior. Detection of object's certain behaviors in the crowd has become a critical problem because in the event of terror strikes. Recognition of object's certain behaviors is an important but difficult problem in the area of computer vision. As the realization of big data utilizing machine learning, data mining techniques, the amount of video through the CCTV, Smart-phone and Drone's video has increased dramatically. In this paper, we propose a multiple-sliding window method to recognize the cumulative change as one piece in order to improve the accuracy of the recognition. The experimental results demonstrated the method was robust and efficient learning units in the classification of certain behaviors.
A few bioinformatics tools have been used to find out conserved regions as probes. We have develo... more A few bioinformatics tools have been used to find out conserved regions as probes. We have developed a system based on a heuristic method with web interfaces to find out conserved regions against microbial genomes. The system runs in real time by using relative entropy in limited narrow regions and detecting similar regions between pair regions with local alignment. The system could be useful to find out conserved regions as genome-wide scale.
In the field of computer graphics, fluid-like motion is an important and challenging problem with... more In the field of computer graphics, fluid-like motion is an important and challenging problem with many applications. Currently, the graphical performance and resolution of the mobile handset is highly improved; therefore, mobile 3D games and physically based animations are also more and more popular items in the mobile services. Yet interactive physically-based fluid simulation is still challenging for mobile games. We studied and implemented physically-based models based on Navier-Stokes equation for fire and smoke effects on mobile 3D games. To make an efficient fire effects on mobile games, we implemented billboard based 2D fluid simulation and gradient based coloring effects. We also implemented external forces on billboard for interactive fire simulation with game characters. The mobile platform of our system is WIPI, which is the standard mobile platform in Korea and NF3D for our 3D programming API. All the simulations have also been adopted for our mobile 3D game, "Rupee Story".
ABSTRACT In this paper, we propose an eecient t e c hnique for the animation of exible objects. M... more ABSTRACT In this paper, we propose an eecient t e c hnique for the animation of exible objects. Mass-spring model was employed to represent the exible objects. The easiest approach to animation with mass-spring model is explicit Euler method, but the explicit Euler method has serious disadvantage that it suuers froinstability problem'. The implicit integration method is a possible solution to overcome the instability problem. However, the most critical aw of the implicit method is that it involves a large linear system. This paper presents a fast animation technique for mass-spring model with approximate implicit method. The proposed technique stably updates the state of n mass-points in O(n) time when the number of total springs are O(n). In order to increase the eeciency of simulation or reduce the numerical errors of the proposed approximate implicit method, the number of mass-points must be as small as possible. However, the coarse discretization with a small number of mass-points generates unrealistic appearance of cloth model. We propose a n e w t e c hnique that generates realistic details of the cloth model even though a small number of mass-points are used for simulation by i n troducing wrinkled cubic spline curve. Key words: cloth animationn mass spring modell implicit methodd realistic detaill wrinkled curve.
Beyond a simple voice service, the advances in performance and resolution of mobile handsets prov... more Beyond a simple voice service, the advances in performance and resolution of mobile handsets provides us various features in the field of mobile services. Also, advances in the human-machine interface, electronics and embedded system technologies have created the foundation for a dramatic shift in the way the vehicles are diagnosed and maintained. In this paper, we will show the vehicle diagnostic system using the mobile handset based on the OBD-II (On Board Diagnostics version II) protocol. After 2005, OBD-II standard is obligatory for all vehicles selling in Korea. Our research topics are OBD-II protocol converter for the mobile handset and the vehicle diagnostics system on WIPI(Wireless Internet Platform for Interoperability) mobile platform, which is the standard mobile platform in Korea.
Interactive and realistic rendering of various effects which are appeared in the real world paint... more Interactive and realistic rendering of various effects which are appeared in the real world paintings can be used for several applications such as animation or education. For Western paintings, watercolor or oil paintings have been main subjects for representation and model-based approaches have been researched. In case of Oriental paintings, numerous studies have focused on model-based approaches. In this paper, a new model-based approach for Oriental paintings is described. To represent the diffusion effects, we propose local equilibrium model(LEM) which is a method to calculate the movement of water and ink effectively. With LEM, the diffusion of water and ink can be processed interactively and realistic diffusion effects can be generated. In addition, we propose a layer model. With this layer model, overlapped strokes can be well represented. Conclusively, we show several results of Oriental paintings drawn with proposed methods.
Interactive and realistic rendering of various effects which are appeared in the real world paint... more Interactive and realistic rendering of various effects which are appeared in the real world paintings can be used for several applications such as animation or education. For Western paintings, watercolor or oil paintings have been main subjects for representation and model-based approaches have been researched. In case of Oriental paintings, numerous studies have focused on model-based approaches. In this paper, a new model-based approach for Oriental paintings is described. To represent the diffusion effects, we propose local equilibrium model(LEM) which is a method to calculate the movement of water and ink effectively. With LEM, the diffusion of water and ink can be processed interactively and realistic diffusion effects can be generated. In addition, we propose a layer model. With this layer model, overlapped strokes can be well represented. Conclusively, we show several results of Oriental paintings drawn with proposed methods.
In this paper, we propose a novel extracting method which is to estimate the 3D coordinates of th... more In this paper, we propose a novel extracting method which is to estimate the 3D coordinates of the finger joints based on markless motion capture using forward and inverse kinematics which are used widely in robotics and computer animation. The proposed method estimates finger joints position using IK method in finger joint model made by recognized fingertip and constraints of finger movement. For improvement of extracting finger joint method in folded finger, we propose a new method to estimate the hidden finger joints using known joint position and range of finger movement. The proposed method expected to apply to UI/UX design for user convenience. Keywords: hand gesture, inverse kinematics, depth camera, CCD ․본 연구는 부산대학교 자유과제 학술연구비(2년) 지원에 의해 수행되었음
Next Generation Sequencing(NGS) is a faster and lower cost new sequencing technique getting huge-... more Next Generation Sequencing(NGS) is a faster and lower cost new sequencing technique getting huge-large scale fragment, called reads, sequence data than traditional sequencing. The next step to analyze the large-scale reads is to align to the reference genome sequence which is a fundamental and most important stage. Alignment tools in NGS are used for mapping short reads quickly using a indexed reference sequence. Most general alignment tools process listed reads sequentially without analyzing redundancy of inter-read of NGS data. In this paper, we explore efficient common subsequences using redundancy of inter-read in characteristics of NGS data. Relation between redundant common subsequences and reads containing them is modeled as hitting set problem in graph theory. Also we suggest a way to improve performance of alignment stage.
There are many types of data base for biological and medical data and some of them are connected ... more There are many types of data base for biological and medical data and some of them are connected each other for processing. That is a small information comparing to potential one of them. The study of association in same data base and similar data have been for a long time and now new approaches have announced for integrated data and system because it give us much information than one from simple connection. We have proposed a tool, Evoltap, to integrate different types of data bases and to visualize the relation of attributes. Evoltap is based on an integrative relation graph of an attributes between heterogeneous databases. The program uses 14 databases which are disease, drug, pathway, gene and protein database. It shows the relation graph which consists node for instance of attribute and link for association of the instance directly or indirectly in 14 databases. Evoltap mines a new relation from the combining the data bases in silico instead of trials in vitro. It also has capa...
2008 IEEE 8th International Conference on Computer and Information Technology Workshops, 2008
CINETOON: A Semi-Automated System for Rendering Black/White Comic books from Video Streams Dong-S... more CINETOON: A Semi-Automated System for Rendering Black/White Comic books from Video Streams Dong-Sung Ryu , Soo-Hyun Park , Jeong-won Lee , Do-Hoon Lee and Hwan-Gue Cho Dept. ... Therefore those of hard-trained artists can draw the commercial comics with easily. ...
Proceedings of the 9th International Conference on Ubiquitous Information Management and Communication - IMCOM '15, 2015
A time series is a sequence of data sequentially observed over time. It contains important inform... more A time series is a sequence of data sequentially observed over time. It contains important information such as stock, futures, foreign exchange rate, brainwave, heart rate, weather change, wind speed, air temperature and advective flux. Analyzing time series data is essential to estimate the future and understand scientific characteristics inherent in many phenomenon. There are a large number of methods for analyzing individual time series, especially on similarity measures. However, there is a lack in considering two time series simultaneously and discovering their correlation. In this paper, we propose a pairwise time-series walk plot model, which analyzes the correlation between two time series and visualizes both of them at the same time. And then we present a measure that represents the correlation of two time series by quantifying their complicated boundary shape.
Journal of the Korea Academia-Industrial cooperation Society, 2012
There are still problems have to solve for safety of driving comparing to the upgraded convenienc... more There are still problems have to solve for safety of driving comparing to the upgraded convenience of advanced vehicle. Most traffic accident is by uncareful driving cause of interface operations which are directive reasons of it in controlling the complicate multimedia device. According to interesting in smart automobile, various approaches for safe driving have been studied. The current multimedia interface embedded in vehicle is lacking the safety due to loss the sense and operation capacity by instantaneous view movement. In this paper, we propose a safe dialing system for safe driving to control dial and search dictionary by hand-gesture. The proposed system improved the user convenience and safety in automobile operation using intuitive gesture and TTS(Text to Speech).
Human action recognition is an active topic in computer vision. In recent years, tangible results... more Human action recognition is an active topic in computer vision. In recent years, tangible results have been shown through deep learning methods; however, at a very high computational cost. They may be suitable for video retrieval or video summarization of a movie or drama, but they are not suitable for the visual surveillance of human action, which should be analyzed in real time. In this study, we propose an action change detection method to reduce the computational cost. On the surveillance camera screen, anomalous actions such as movies and sports are frequently not observed, and simple actions are often repeated. Thus, it is very inefficient to continue to apply high cost action recognition methods on repeated actions. The proposed action change detection method decides whether the previous action of the person is the same as the current action. The action recognition method is applied only when it has determined that the action has changed. The action change detection process i...
Abstract In pattern recognition, the classification accuracy has a strong correlation with the se... more Abstract In pattern recognition, the classification accuracy has a strong correlation with the selected features. Therefore, in the present paper, we applied an evolutionary algorithm in combination with linear discriminant analysis (LDA) to enhance the feature selection in a static image-based facial expressions system. The accuracy of the classification depends on whether the features are well representing the expression or not. Therefore the optimization of the selected features will automatically improve the classification accuracy. The proposed method not only improves the classification but also reduces the dimensionality of features. Our approach outperforms linear-based dimensionality reduction algorithms and other existing genetic-based feature selection algorithms. Further, we compare our approach with VGG (Visual Geometry Group)-face convolutional neural network (CNN), according to the experimental results, the overall accuracy is 98.67% for either our approach or VGG-face. However, the proposed method outperforms CNN in terms of training time and features size. The proposed method proves that it is able to achieve high accuracy by using far fewer features than CNN and within a reasonable training time.
Behavior of intelligent video surveillance system is recognized by analyzing the pattern of the o... more Behavior of intelligent video surveillance system is recognized by analyzing the pattern of the object of interest by using the frame information of video inputted from the camera and analyzes the behavior. Detection of object's certain behaviors in the crowd has become a critical problem because in the event of terror strikes. Recognition of object's certain behaviors is an important but difficult problem in the area of computer vision. As the realization of big data utilizing machine learning, data mining techniques, the amount of video through the CCTV, Smart-phone and Drone's video has increased dramatically. In this paper, we propose a multiple-sliding window method to recognize the cumulative change as one piece in order to improve the accuracy of the recognition. The experimental results demonstrated the method was robust and efficient learning units in the classification of certain behaviors.
A few bioinformatics tools have been used to find out conserved regions as probes. We have develo... more A few bioinformatics tools have been used to find out conserved regions as probes. We have developed a system based on a heuristic method with web interfaces to find out conserved regions against microbial genomes. The system runs in real time by using relative entropy in limited narrow regions and detecting similar regions between pair regions with local alignment. The system could be useful to find out conserved regions as genome-wide scale.
In the field of computer graphics, fluid-like motion is an important and challenging problem with... more In the field of computer graphics, fluid-like motion is an important and challenging problem with many applications. Currently, the graphical performance and resolution of the mobile handset is highly improved; therefore, mobile 3D games and physically based animations are also more and more popular items in the mobile services. Yet interactive physically-based fluid simulation is still challenging for mobile games. We studied and implemented physically-based models based on Navier-Stokes equation for fire and smoke effects on mobile 3D games. To make an efficient fire effects on mobile games, we implemented billboard based 2D fluid simulation and gradient based coloring effects. We also implemented external forces on billboard for interactive fire simulation with game characters. The mobile platform of our system is WIPI, which is the standard mobile platform in Korea and NF3D for our 3D programming API. All the simulations have also been adopted for our mobile 3D game, "Rupee Story".
ABSTRACT In this paper, we propose an eecient t e c hnique for the animation of exible objects. M... more ABSTRACT In this paper, we propose an eecient t e c hnique for the animation of exible objects. Mass-spring model was employed to represent the exible objects. The easiest approach to animation with mass-spring model is explicit Euler method, but the explicit Euler method has serious disadvantage that it suuers froinstability problem'. The implicit integration method is a possible solution to overcome the instability problem. However, the most critical aw of the implicit method is that it involves a large linear system. This paper presents a fast animation technique for mass-spring model with approximate implicit method. The proposed technique stably updates the state of n mass-points in O(n) time when the number of total springs are O(n). In order to increase the eeciency of simulation or reduce the numerical errors of the proposed approximate implicit method, the number of mass-points must be as small as possible. However, the coarse discretization with a small number of mass-points generates unrealistic appearance of cloth model. We propose a n e w t e c hnique that generates realistic details of the cloth model even though a small number of mass-points are used for simulation by i n troducing wrinkled cubic spline curve. Key words: cloth animationn mass spring modell implicit methodd realistic detaill wrinkled curve.
Beyond a simple voice service, the advances in performance and resolution of mobile handsets prov... more Beyond a simple voice service, the advances in performance and resolution of mobile handsets provides us various features in the field of mobile services. Also, advances in the human-machine interface, electronics and embedded system technologies have created the foundation for a dramatic shift in the way the vehicles are diagnosed and maintained. In this paper, we will show the vehicle diagnostic system using the mobile handset based on the OBD-II (On Board Diagnostics version II) protocol. After 2005, OBD-II standard is obligatory for all vehicles selling in Korea. Our research topics are OBD-II protocol converter for the mobile handset and the vehicle diagnostics system on WIPI(Wireless Internet Platform for Interoperability) mobile platform, which is the standard mobile platform in Korea.
Interactive and realistic rendering of various effects which are appeared in the real world paint... more Interactive and realistic rendering of various effects which are appeared in the real world paintings can be used for several applications such as animation or education. For Western paintings, watercolor or oil paintings have been main subjects for representation and model-based approaches have been researched. In case of Oriental paintings, numerous studies have focused on model-based approaches. In this paper, a new model-based approach for Oriental paintings is described. To represent the diffusion effects, we propose local equilibrium model(LEM) which is a method to calculate the movement of water and ink effectively. With LEM, the diffusion of water and ink can be processed interactively and realistic diffusion effects can be generated. In addition, we propose a layer model. With this layer model, overlapped strokes can be well represented. Conclusively, we show several results of Oriental paintings drawn with proposed methods.
Interactive and realistic rendering of various effects which are appeared in the real world paint... more Interactive and realistic rendering of various effects which are appeared in the real world paintings can be used for several applications such as animation or education. For Western paintings, watercolor or oil paintings have been main subjects for representation and model-based approaches have been researched. In case of Oriental paintings, numerous studies have focused on model-based approaches. In this paper, a new model-based approach for Oriental paintings is described. To represent the diffusion effects, we propose local equilibrium model(LEM) which is a method to calculate the movement of water and ink effectively. With LEM, the diffusion of water and ink can be processed interactively and realistic diffusion effects can be generated. In addition, we propose a layer model. With this layer model, overlapped strokes can be well represented. Conclusively, we show several results of Oriental paintings drawn with proposed methods.
In this paper, we propose a novel extracting method which is to estimate the 3D coordinates of th... more In this paper, we propose a novel extracting method which is to estimate the 3D coordinates of the finger joints based on markless motion capture using forward and inverse kinematics which are used widely in robotics and computer animation. The proposed method estimates finger joints position using IK method in finger joint model made by recognized fingertip and constraints of finger movement. For improvement of extracting finger joint method in folded finger, we propose a new method to estimate the hidden finger joints using known joint position and range of finger movement. The proposed method expected to apply to UI/UX design for user convenience. Keywords: hand gesture, inverse kinematics, depth camera, CCD ․본 연구는 부산대학교 자유과제 학술연구비(2년) 지원에 의해 수행되었음
Next Generation Sequencing(NGS) is a faster and lower cost new sequencing technique getting huge-... more Next Generation Sequencing(NGS) is a faster and lower cost new sequencing technique getting huge-large scale fragment, called reads, sequence data than traditional sequencing. The next step to analyze the large-scale reads is to align to the reference genome sequence which is a fundamental and most important stage. Alignment tools in NGS are used for mapping short reads quickly using a indexed reference sequence. Most general alignment tools process listed reads sequentially without analyzing redundancy of inter-read of NGS data. In this paper, we explore efficient common subsequences using redundancy of inter-read in characteristics of NGS data. Relation between redundant common subsequences and reads containing them is modeled as hitting set problem in graph theory. Also we suggest a way to improve performance of alignment stage.
There are many types of data base for biological and medical data and some of them are connected ... more There are many types of data base for biological and medical data and some of them are connected each other for processing. That is a small information comparing to potential one of them. The study of association in same data base and similar data have been for a long time and now new approaches have announced for integrated data and system because it give us much information than one from simple connection. We have proposed a tool, Evoltap, to integrate different types of data bases and to visualize the relation of attributes. Evoltap is based on an integrative relation graph of an attributes between heterogeneous databases. The program uses 14 databases which are disease, drug, pathway, gene and protein database. It shows the relation graph which consists node for instance of attribute and link for association of the instance directly or indirectly in 14 databases. Evoltap mines a new relation from the combining the data bases in silico instead of trials in vitro. It also has capa...
2008 IEEE 8th International Conference on Computer and Information Technology Workshops, 2008
CINETOON: A Semi-Automated System for Rendering Black/White Comic books from Video Streams Dong-S... more CINETOON: A Semi-Automated System for Rendering Black/White Comic books from Video Streams Dong-Sung Ryu , Soo-Hyun Park , Jeong-won Lee , Do-Hoon Lee and Hwan-Gue Cho Dept. ... Therefore those of hard-trained artists can draw the commercial comics with easily. ...
Proceedings of the 9th International Conference on Ubiquitous Information Management and Communication - IMCOM '15, 2015
A time series is a sequence of data sequentially observed over time. It contains important inform... more A time series is a sequence of data sequentially observed over time. It contains important information such as stock, futures, foreign exchange rate, brainwave, heart rate, weather change, wind speed, air temperature and advective flux. Analyzing time series data is essential to estimate the future and understand scientific characteristics inherent in many phenomenon. There are a large number of methods for analyzing individual time series, especially on similarity measures. However, there is a lack in considering two time series simultaneously and discovering their correlation. In this paper, we propose a pairwise time-series walk plot model, which analyzes the correlation between two time series and visualizes both of them at the same time. And then we present a measure that represents the correlation of two time series by quantifying their complicated boundary shape.
Journal of the Korea Academia-Industrial cooperation Society, 2012
There are still problems have to solve for safety of driving comparing to the upgraded convenienc... more There are still problems have to solve for safety of driving comparing to the upgraded convenience of advanced vehicle. Most traffic accident is by uncareful driving cause of interface operations which are directive reasons of it in controlling the complicate multimedia device. According to interesting in smart automobile, various approaches for safe driving have been studied. The current multimedia interface embedded in vehicle is lacking the safety due to loss the sense and operation capacity by instantaneous view movement. In this paper, we propose a safe dialing system for safe driving to control dial and search dictionary by hand-gesture. The proposed system improved the user convenience and safety in automobile operation using intuitive gesture and TTS(Text to Speech).
Uploads
Papers by DoHoon Lee