Ryuichi Oka
Professor
Mohamed Hamada
Associate Professor
Keitaro Naruse
Assistant Professor
Yu Nakajima
Assistant Instructor
|
Ryuichi Oka Professor |
Mohamed Hamada Associate Professor |
Keitaro Naruse Assistant Professor |
Yu Nakajima Assistant Instructor |
First and foremost, the Image Processing Laboratory engages in research and development of image-based pattern recognition including some areas of Artificial Intelligence, database organisation, retrieval, and robotics. More specifically, as can be seen from the background information and the recent research publications of the members of the laboratory, 2-dimensional Continuous Dynamic Programming for non-linear image matching, 3D image reconstruction, object tracking in a time-varying image, image retrieval from video data, and extracion of the Internet community are our current focus.
Related to the recent progress of the web, a huge amount of multimedia data without index becomes available to store in our PC's. However, no sophistcated methodology to manage such data has been developed so far. So that we are solicited to attach an index to each data. Our research aim is to develop algorithms to realize automatic annotation to real word data for integrated retrieval of multimedia information. The algorithms include self-organisation and transformation among representation of multimedia and feature extraction and recognition of real data. Real world data includes video, still image, speech, music, sound, and text each of which has not been indexed by labels.
An example of multimedia integration retrieval called "CrossMediator" developed by ten-year project (1992-2002) RWC of METI (Japan). Some parts of CrossMediator have been in the commercial market by through a private company. Prof. R.Oka was a chief of group which developed "CrossMediator". Our laboratory will pursue to develop more sophisticated functions which might reveal a new generation of the Internet. Speech retreival system used in CrossMediator is used for developing new types of commercial products. At the present time we use two kinds of index for searching video data. One is used to obtain the target media from Youtube by using index. The other one is used for searching parts of the target data which are not indexed.
One of the recent progress is developing a set of algorithms for spotting recognition and tracking objects in a time-varying image. These algorithms enable to realize automatic anonotation of video image capturing moving objects. The most promissing algorithm called 2-dimensional Continuous Dynamic programming (2DCDP) was proposed by our laboratory. This enables to perform full pixel matching between images. The full pixel correspondene is used for reconstructing 3D shape from at least 3 images obtained from different viewpoints and also gives a quite precise features like optical low from a motion image. The 2DCDP is also applicable to non-linear registration for medical images to extract different parts between normal CT imagesand abnormal CT images.
Another recent progress is developing a spotting algorithm for song wave retrieval. A large amount of song wave data is available for anyone along with the progress of storage hardware of music data. But the progress of technology for accessing the song wave data seems quite slow so that we must use conventional and not sophistcated tools. Our new software can provide users a convenient way to find and rerieve the song wave data by singing an arbitrary part of it which he/she wants to hear.
The Internet is regarded as a network composed of virtual communities. Visualization of the community is becoming an important research target. Our lab is developing a sophistcated algorithm based on so-called Associated hyper-linked word space (A-space). We use the algorithm to detect clusters each of which corresponds a virtual community. Visualization of each virtual community is realized by showing the content extracted from the web pages included the community.
On the other hand, we have carried out several research projects related to complex systems in which many independent components interact with each other and emerge global phenomena. Particularly, we have conducted on multiple robot control, and energy efficient walking control in biped robots.
With respect to multiple robot control, we have analyzed the dynamics of multiple mobile robots as a swarm under the mixture condition of long and short range communication, and evaluated the required amount of communication for the multiple robot control. Particularly, we have investigated the relation between the speed of convergence and the graph spectra induced from the swarm robot network, and we have developed a stable and fast swarm robot control method. As implementing actual multiple mobile robots, we need a sensing system. In this project, we have developed a sensing and position estimating system for the robots, and evaluated the accumulated errors.
In terms of the walking control of biped robots, we have set the problem to walk down steep hills, which is known as steps of walking robots fall into chaotic and unstable one. We have developed a control method for the problem, and evaluated it with numerical experiments. The results show it is capable to walk down the steep slopes stably with less control energy consumption. is capable to walk down the steep slopes stably with less control energy consumption.
Mohamed Hamada and Sayota Sato. Simulator and Robot-Based Game for Learning Automata Theory. Lecture Notes Computer Science, 6249(2010):428-436, 2010.
This paper introduces a Finite State Machine Simulator and a robotbased game associated with it for learning Automata Theory.
Mohamed Hamada and Kosuke Nishikawa. Web-Based Enhanced Learning Style Index with Integration into an e-Learning System. Lecture Notes in Computer Science, 6483(2010):101-110, 2010.
This paper is concerned on the study, implementation, and application of a web-based learning style index. We also show a case study on the integration of learning style index into an e-learning system.
H. El-Bakry and Mohamed Hamada. A Novel Watermark Technique for Relational Databases. Lecture Notes in Computer Science, 6320(2010):226 232, 2010.
In this paper, a new approach for protecting the ownership of relational database is presented.
Mohamed Hamada and Sayota Sato. Lego NXT as a learning tool. In John Impagliazzo Reyyan Ayfer and Editors Cary Laxer, editors, ACM ITiCSE 2010, pages 321-323, New York, June 2010. ACM, ACM.
The paper introduces a novel idea to use NXT Lego set as a learning tool
K. Naruse, K. Suenaga, and S. Hukui. Control and Localization in Swarm Leading Model. In Proc. of The 14th Asia Pacific Symposium on Intelligent and Evolutionary Systems, pages CDROM, 2010.
The objective of this paper is to solve the dynamic plane covering problem by the movement of multiple robots. One of the ways to solve the problem is the swarm leading control method, in which one robot, called a leader, gives a motion of the robots, and all the other robots, called followers, move around the leader with a local measurement and control. However, because the robots suffer from noisy measurement, the stability of the swarm being connected is unknown. In this paper, the stability conditions are be shown and proved. The proposed control method is verified with computational simulations considering the sensor noises.
S. Hukui, J. Luo, T. Daicho, K. Suenaga, and K. Naruse. Mutual Localization of Multiple Sensor Node Robots. In Proc. of 2nd International Symposium on Aware Computing, pages CDROM, 2010.
The objective of this paper is to develop a localization system for cooperative multiple mobile robots, in which each robot is assumed to observe a set of known landmarks and equipped with an omnidirectional camera. If only a limited number of landmarks are available, it suffers from law accuracy. In this paper, it is assumed that a robot can detect other robots by using the omnidirectional camera and know their positions estimated by themselves and utilize the positions for its localization. In other words each robot can be viewed as a new mobile landmark. However, since the position information estimated by the robot itself is inaccurate, the estimated position of the robot cannot be high enough. This paper presents an investigation of the self localization error of a robot using Extended Kalman Filter to solve the localization problem with the insufficient landmarks and inaccurate position information.
K. Naruse, K. Suenaga, and S. Hukui. Dynamic Plane Covering by Movement of Swarm. In Proc. of Joint 5th International Conference on Soft Computing and Intelligent Systems and 11th International Symposium on Advanced Intelligent Systems, pages CDROM, 2010.
The objective of this paper is to solve the dynamic plane covering problem by the movement of multiple robots, such as sprinkling water to a large field by several vehicles or aircrafts. In the problem, all of the points in the field should be covered with an equal density. One of the ways to solve the problem is the swarm leading control method, in which one robot, called a leader, moves along a path in the field, and all the other robots, called followers, move around the leader with a fixed distance. However, because the leader and the followers have asymmetrical relations, the stability condition of the swarm leading control method is unknown. In this paper, introducing a Manhattan norm for distance measurement, the stability condition can be shown and proved. The proposed control method is verified with computational simulations.
K. Naruse and K. Suenaga. Stability Analysis of Multiple Mobile Robots by Swarm Leading Control. In Proc. of International Conference on Advanced Mechatronics, pages CDROM, 2010.
This paper presents a new control model called the swarm leading for navigating multiple mobile robots as a swarm. In the control model, only a single robot, called a leader, knows a destination of the swarm, all the other robots, called followers, determine their motion by interacting with neighbor robots. If the motion of the leader is propagated to all the followers, the swarm is navigated. However, if some of the followers are left away from the other robots and isolated, they will be never navigated. Therefore, we need a stable control model that all the robots are navigated. In this paper, we develop the swarm leading control model and show the conditions that all the robots are controllable analytically using the algebraic graph theory. The proposed control model is verified with numerical experiments.
Jun Ma, Long Zheng, Yuichi Yaguchi, Mianxiong Dong, and Ryuichi Oka. Image Classification Based on segmentation-free Object Recognition. In IEEE 17-th ICIP, pages pp.2157-2160, 2010.
This paper presents a new method for categorical classification. A method called two-dimensional continuous dynamic programming (2DCDP) is adopted to optimally capture the corresponding pixels within nonlinearly matched areas in an input image and a reference image representing an object without advance segmentation procedure. Then an image can be converted into a direction pattern which is made by matching pixels between a reference image and an input image. Finally, the category of the test image is deemed to be that which has the strongest correlation with the learning images. Experimental results show that the proposed method achieves a competitive performance on the Caltech 101 image dataset.
Hisato Aota, Kazuhiro Ota, Yuichi Yaguchi, and Ryuichi Oka. Extracting Objects by Clustering of Full Pixel Trajectories. In SIGMAP 2010, pages pp.65-72, 2010.
The purpose of this paper is to propose a novel method for both segmentation of objects and extraction of motion features for moving objects in a video data. The method adopts an algorithm called two-dimensional dynamic programming (2DCDP) for extracting pixel-wise trajectories. A clustering algorithm is applied to a set of pixel trajectories to determine objects each of which is corresponding to a trajectory cluster. We conduct experiments for comparing our method with the conventional methods such as KLT trackeras SIFT. The experiment shows that our method is more powerful than the conventional methods.
Yukihiro Yoshida, Koushi Yamaguchi abd Yuichi Yaguchi, Yuichi Okumura, Ken ichi Kuroda, and Ryuichi Oka. Accelerate Two-dimensional Continuous Dynamic Programming by memory Reduction and Parallel Processing. In International Conference Applied Computing 2010, pages pp.61-68, Oct. 2010.
This paper contains a proposal for optimizing and accelerating the computation of two-dimensional continuous dynamic programming (2DCDP). 2DCDP processing is optimized by memory reduction and parallelization using OpenMP. We apply buffer resizing and utilize toggle-type buffers to reduce the required memory size. In addition, same-rank processes and pixel correspondence calculation are parallelized by OpenMP instructions to reduce the computation cost/time of 2DCDP. For accumulation, we also apply a realignment of buffering addresses for SIMD on multi-cores/multi-processors. The experimental results show that the computational time and the memory usage have reduced to about 1/4 and 1/5 of the original ones, respectively. Moreover, the concurrency of 2DCDP hot-spot is improved from 5.8 to 7.1 on a quad-core CPU with 8 threads.
K. Suenaga and K. Naruse. Study of Swarm-Leading Control Model with 1-norm. In 2010 FAN Symposium, 2010.
K. Tamura and F. Hukui annd K. Naruse. Mutual Localization and Position Control by Swarm Robots. In 2010 SICE SI Division Conference, 2010.
K. Suenaga and K. Naruse. Consensus Problem of Velocity in MultiAgent System. In 2010 Robotics Mechatronics Conference, 2010.
K. Suzuki, T. Suzuki, and K. Naruse. Robustness of Passive Walking in Steep Slope. In 2010 Fall JSPE Conference, 2010.
T. Daicho and K. Naruse. Landmark Recognition in Omnidirectional Vision by High Order Local Autocorrelation Features. In 2010 Spring JSPE Conference, 2010.
T. Sato, K. Suenaga, and K. Naruse. Swarm connected and fragmented condition in obstacle avoidance. In 2010 SICE SI Division Conference, 2010.
J. Luo, S. Hukui, T. Daicho, K. Suenaga, and K. Naruse. Localization of Cooperative Multiple Mobile Robots by Extended Kalman Filter. In 2010 FAN Symposium, 2010.
Ryuichi Oka, Yuichi Yaguchi, and Shinya Mizoe. General Scheme of Continuous Dynamic Programming Optimal Full Pixel Matching for Spotting Image. In IEICE technical Report(2010-09), pages PRMU2010-87 IBISML2010-59, 2010-09.
Two-dimensional Continuous Dynamic Programming (2DCDP) for optimal full pixel matching is discussed. There is a family of 2DCDP algorithms each of them takes a position in a general scheme of CDP. The scheme includes also algorithms for full element matching of 1D-1D, 1D-2D and 3D-3D. Two images to be matched have different image sizes. In order to solve this problem, 2DCDP is designed to conduct the matching in the style of segmentation-free registration of the smaller size reference image on the larger size input image. 2DCDP is a promising algorithm for applying to the applications such as free-viewpoint TV, object tracking, object recognition, optical flow detection etc.
Mohamed Hamada, April 2010.
Member, IEEE
Mohamed Hamada, April 2010.
Member, ACM
K. Naruse, 2010.
Program Committee, IEEE CIT Conference
K. Naruse, 2010.
Program Committee, IEEE ISAC
K. Naruse, 2010.
Organized Session Committee, FAN Symposium, SOFT
K. Naruse, 2010.
Session Organizer, SCIS-ISIS Conference, SICE
K. Naruse, 2010.
Local Arrangement Committee, Robotics Mechatronics Conference, JSME
K. Naruse, 2010.
Program Committee, FAN Symposium, SOFT
Ryuichi Oka, May 2010.
Reviewer of submitted papers, The Institute of Electronics, Information and Communication Engineers.
Ryuichi Oka, May 2010.
Reviewer of submitted papers, The Acoustic Society of Japan.
Ryuichi Oka, April 2010.
Senior Member of editorial committee, Japanese Society for Artificial Intelligence.
Ryuichi Oka, April 2010.
Reviwer of submitted papers, The nine Inter. Conf. on Computer and Information Technology.
Takayuki Hoshi. Graduation Thesis: Application of Study on Learning Preferences, School of Computer Science and Engineering, March 2011.
Thesis Adviser: M. Hamada
Hayato Namae. Graduation Thesis: A Learning System for some Audio Algorithms, School of Computer Science and Engineering, March 2011.
Thesis Adviser: M. Hamada
Kazuma Suzuki. Graduation Thesis: Robustness of Semi-Passive Dynamic Walking Models for Steep Slopes, School of Computer Science and Engineering, March 2011.
Thesis Adviser: Keitaro Naruse
Kazuya Tamura. Graduation Thesis: Investigation of Relations between Multi Robot Localization and the Stability, School of Computer Science and Engineering, 2010.
Thesis Adviser: Keitaro Naruse
Takayuki Suzuki. Graduation Thesis: Highly Efficient Design for SemiPassive Dynamic Walking Models on Level Ground, School of Computer Science and Engineering, March 2011.
Thesis Adviser: Keitaro Naruse
Tatsuya Sato. Graduation Thesis: Analysis of Swarm Connected and Fragmented Condition in Obstacle Avoidance, School of Computer Science and Engineering, 2010.
Thesis Adviser: Keitaro Naruse
Jie Luo. Master Thesis: Multiple Robot Simultaneous Localization Considering Other Robots as Mobile Landmarks, Graduate School of Computer Science and Engineering, September 2010.
Thesis Adviser: Keitaro Naruse
Keigo Suenaga. Master Thesis: Analysis of Linear Swarm Control Methodby Algebraic Approach, Graduate School of Computer Science and Engineering, March 2011.
Thesis Adviser: Keitaro Naruse
2011 THE UNIVERSITY OF AIZU ALL RIGHTS RESERVED