| 1. |
階層型情報構造化環境プラットフォームの開発. |
| 2. |
Person recognition based on shadows. |
| 3. |
第四人称視点に基づく情報構造化空間の状況説明文生成. |
| 4. |
Gait recognition with low resolution images using inter-frame difference and CNN
Gait-based person recognition has received an increasing amount of attentions for monitoring and surveillance applications. One of issues in gait recognition is that it is difficult to recognize people with high performance, in case that the resolution of captured images is too low. To deal with this problem, this paper presents SFDEINet, which uses Singed Frame Difference Energy Image (SFDEI) as input images. SFDEI has an advantage of explicit representation of motion of walking people, by changing the size of frame difference. To take the size of frame difference into account, SFDEINet is deigned to have convolution layers in parallel, followed by fusing outputs of each layer. We will show the SFDEINet is more suitable for SFDEI than GEINet in experiments. . |
| 5. |
生活支援サービスのための移動型インフレータブルロボットアームの開発. |
| 6. |
深層学習と術具3次元形状モデルの組み合わせによるロボット支援内視鏡手術画像からの術具位置姿勢推定. |
| 7. |
Development of Automatic Driving System for Personal Mobility Vehicles using Small Sensor Terminal
In recent years, Personal Mobility Vehicle (PMV) have been attracting much attention as the new future vehicle. In this paper, we propose a new control system which provides automatic driving function or driving support function to PMVs using a small sensor terminal named "Portable". This system enables to add intelligent functions to various PMVs by attaching the Portable. In this paper we introduce two types of intelligent PMVs, standing type mobility and electric wheelchair, and show experimental results in indoor and outdoor environments including a theme park. . |
| 8. |
Development of informationally structured environment for personal mobility and navigation experiment at theme park
This paper proposes distributed sensor systems to construct an informationally structured environment by using LRFs (Laser Range Finders) and active beacons. With the proposed system, a personal mobility can localize its position and automatically move in indoor and outdoor environments. Experiments show that the personal mobility is guided appropriately in various environments, such as the campus of the university, the exhibition booth, and the theme park. . |
| 9. |
テーマパークにおける自律案内ロボットの開発. |
| 10. |
Realization of Pick and Place Operations by an Inflatable Robotic Arm
This paper presents an overview of pick and place operations performed by an inflatable robot. The inflatable robotic arm is composed of inflatable links and pneumatic bag actuators. This technology is expected to be applied to service robots working in human daily life, since the robotic arm has many advantages such as lightness, softness and safety. The pick and place tasks are realized by both pressure and visual feedback controls. The usefulness of the robot is demonstrated by several real world experiments. . |
| 11. |
Development of Tour-Guide Robot System at Theme Park using Quasi-Zenith Satellite System
Over the past decades, 3D jobs (dangerous, dirty, demanding) are expected to be replaced by autonomous robots. At a theme park, surveillance, cleaning, and guiding tasks can be regarded as 3D jobs. This paper aims to develop an autonomous tour guide robot system at a large theme park. One of the characteristics of the tour guide robot we developed is the use of QZSS (quasi-zenith satellite system) and CLAS (centimeter level augmentation service), which realizes the centimeter-level positioning for autonomous service robots. We also implemented autonomous tour-guide system including path planning, collision avoidance, and voice explanation. Experimental results at the theme park showed that QZSS has sufficient localization accuracy for the tour-guide task, and the tour-guide experiment was successfully carried out. . |
| 12. |
Development of dementia care education system combining augmented reality and distributed tactile sensor
In this study, we aim to quantify care skills in the care technique called "Humanitude". Humanitude has been attracting much attention as a gentle and effective care technique for a dementia patient in recent years. The developed wearable tactile sensor aims to quantify the "touch" skill among the four representative skills in Humanitude, which are gaze, touch, talk, and stand up. In addition, we developed a learning system of Humanitude using AR technology and a real object, which realizes the interaction between a care-giver and a patient and enabled to quantify the "touch" and "gaze" skill. . |
| 13. |
3D Cell Shape Reconstruction From Multifocal Microscope Image Sequence. |
| 14. |
Development of 3D motion measurement system for sport climbing
Sport climbing is one of the sports attracting attention all over the world and expected to become more popular. Therefore, motion measurement and analysis system have been required to improve the performance of players and instruct them. In several researches, the systems on motion measurement and analysis of climbing have been presented. Most of these systems consider only the movement of climbing, and take no thought of the force on holds. However, force information is indispensable to analyze the body motion of players deeply. This paper proposes a new 3D motion and force measurement system for sport climbing that can measure the movement and the force on holds simultaneously. Then, this proposed system enables us to measure and analyze climbing motions geometrically and dynamically. In order to verify the performance of the proposed system, we conducted the experiment in an actual climbing gym. In addition, we analyzed the center of mass of a player and the force on a hold by using the proposed system. . |
| 15. |
Previewed Reality 2.0: Near-future perception system:-Building system and experiment of collision avoidance with transmission type display-
For safe human-robot coexistence, we have developed a near-future perception system named Previewed Reality. In this paper, we propose the latest system "Previewed Reality 2.0", which realizes a quasi-real-time processing. The system consists of an informationally structured environment (ISE), a Microsoft HoloLens, and a dynamic simulator. In an ISE, a number of sensors are embedded, and information such as the position of furniture, objects, humans, and robots, is sensed and stored in a database. Therefore, we can predict the next possible events using a dynamic simulator and synthesize virtual images of what users will see in the near future from their own viewpoint. The virtual images are presented to the user by overlaid on a real scene using augmented reality technology. The proposed system allows a human to avoid collision with a robot by showing possible hazardous situations to the human intuitively in advance. . |
| 16. |
Development of a Lightweight Soft Gripper using Snap-Through-Buckling: 1st Report Prototype of Actuator Part
As an effective way to grasp an object softly and robustly, many soft grippers have been developed and reported. However, it is still difficult to grasp objects that are easily damaged by local contact forces such as fruits. This paper proposes a soft gripper which grasps such easily damaged objects by wrapping around. This approach gives a wide area of contact and reduces local contact forces. Moreover, this gripper is composed of only soft materials and integrally molded. The opening-closing mechanism is realized by using snap-through-buckling. In this paper, we show the design concept and the performance of the first prototype of the gripper. . |
| 17. |
Development of a Lightweight Soft Gripper using Snap-Through-Buckling: Part 2: Structural Optimization of an Actuator Part. |
| 18. |
Measurement and Assessment of Touch Skills during Dementia Care Movements Using Tactile Gloves. |
| 19. |
Development of Tour-Guide and Co-Experience Robot System at a Theme Park using 5th Generation Mobile Communication System
3D jobs (dangerous, dirty, demanding) are expected to be replaced by autonomous robots. At a theme park, surveillance, cleaning, and guiding tasks can be regarded as 3D jobs. This paper aims to develop an autonomous tour guide robot system and co-experience system at a large theme park. One of the characteristics of the tour guide robot we developed is the use of QZSS (quasi-zenith satellite system) and CLAS (centimeter-level augmentation service), which realizes the centimeter-level positioning for autonomous service robots. On the other hand, co-experience realizes the sharing of experience through the robot utilizing the 5th generation mobile communication system. Experimental results at the theme park showed that the tour-guide experiment was successfully carried out and the co-experience system can provide sharing of the experience of the robot to the user. . |
| 20. |
Development of dementia care training system combining augmented reality and distributed tactile sensor: Scenario-based training system and demonstration
This paper proposes a training system for a multimodal comprehensive care methodology for dementia patients named Humanitude. Humanitude has been attracting much attention as an effective care technique for dementia patients. Humanitude consists of four main techniques, namely, eye contact, verbal communication, touch, and standing up, and more than 150 care elements. Learning Humanitude thus requires much time. To train Humanitude effectively, we develop a training systemthat realizes sensing and interaction simultaneously by combining areal entity and augmented reality technology. To imitate the interaction between a patient and a caregiver, we superimpose a three-dimensional CG model of a patient's face onto the head of a soft doll using augmented reality technology. Touch information such as position and pressure is sensed using a whole body wearable tactile sensor. The effectiveness of the proposed training system is evaluated in public lectures. . |
| 21. |
Development of ROS2-TMS: New Platform for Informationally Structured Environment
We focus on environmental informationally structuring that makes the environment around the robot intelligent, not just the robot itself. We have developed a software platform ROS-TMS to realize a service robot that coexists with humans in an informationally structured environment. ROS is currently being developed with a new ROS2 with significantly changed and improved communication functions and additional internal functions. Therefore, in this research, we develop ROS2-TMS, which adopts ROS2 for ROS-TMS. In this research, we first develop and port various modules, such as a control system for a communication robot. Furthermore, in the development of the voice user interface and task scheduler, not only robots, but also operation commands to various devices that work on the user, are managed. And a new cancellation function is added to stop running tasks, with the aim of improving service security. . |
| 22. |
Development of ROS2-TMS for informationally structured environment: Robot controller modules and cooperative motion of multiple robots. |
| 23. |
Development of wearable whole body tactile sensor for nursing care quantification
In this study, we aim to quantify care skills in the dementia care technique called "Humanitude". Humanitude has been attracting much attention as a gentle and effective care technique for a dementia patient in recent years. The developed wearable whole body tactile sensor aims to quantify the "touch" skill among the four representative skills in Humanitude, which are gaze, touch, talk, and stand up. In addition, we conducted a demonstration experiment using a wearable whole body tactile sensor to quantify nursing care by general caregivers and a Humanitude expert. . |
| 24. |
Kinematic Analysis of Sport Climbing using 3D Motion and Force Measurement System. |
| 25. |
Analysis of Touching Skill in Humanitude-Contact Position with Care Receiver during Posture Change-. |
| 26. |
Development of Visualization Interface of Motion and Force Information for Sport Climbing. |
| 27. |
Development of Outdoor-Surveillance Robot System Using Crawler Robot. |
| 28. |
Development of Smart Previewed Reality-Near Future Perception System using Smartphone-
In this research, we develop a near-future perception system named "Smart Previewed Reality" for humans and robots to coexist safely. This system consists of a smartphone and the management system of informationally structured environment "ROS-TMS". Informationally structured environment is so-called a smart environment, in which various sensors are placed in the space where the robot works, and the information of humans and objects and current and planned robot motion are stored structurally in a database. Therefore, by showing planned robot motion in the database by a smartphone beforehand (Previewed Reality), the user can perceive the near-future robot motion and avoid dangerous situation such as collisions. We develop smartphone applications for Previewed Reality including visualization of robot motion using AR, UI indicators of robot position, and push notification of dangerous situation. . |
| 29. |
People Identification using 3D LiDAR and Long Short-Term Memory Deep Networks. |
| 30. |
3D Brain Atlas Construction from Single Japanese Cadaver Brain. |
| 31. |
Gait-based biometric identification from invisible shadows
This paper presents a shadow-based person identification technique robust to appearance change caused by variations of clothes and carrying conditions. The person identification technique using invisible lights and shadows has many advantages as invisible and undetected sensing. In addition, shadows on the ground carted by multiple lights can be considered as silhouettes captured by virtual cameras at light positions. Therefore, multiple silhouettes are virtually acquired by taking a shadow image with a ceiling camera. However, if the person wears clothes or carries bags which were not appeared in the dataset, the identification performance is aggravated. To deal with this problem, we introduce a new shadow-based identification technique robust to appearance changes. Firstly, we divide a shadow image into several parts, and estimate the discrimination capability for each part based on gait features between gallery datasets and probe dataset. Next, according to the estimated capability, the priorities of these parts for the person identification are controlled adaptively. We constructed a new shadow database with a variety of clothes and bags, and carried out experiments to verify the effectiveness of the proposed technique.. |
| 32. |
Measurement of Moving Objects and Estimation of Human Behavior Using Floor Sensing System in Daily Life Environment
This paper describes a method of measuring moving objects and estimating human behaviors in a room using only one laser range finder (LRF) installed in the room and a strip of mirror attached to a side wall close to a floor. The area of sensing is limited to a plane parallel to and just a few centimeters above the floor, thus covering the whole room with minimal invasion of privacy of a resident while reducing occlusion. The important feature of the measurement consists in processing of both distance and reflectance acquired by the LRF from the surface of the existing objects. This enables immediate distinction of clusters of objects made of different materials in the analysis of the scene cluttered with objects. The human behavior models are effectively utilized to estimate human behavior from LRF data. The experimental results validate the effectiveness of the proposed method.. |
| 33. |
Mapping of Brain Surface Mesh Model onto Arbitrary Closed Surface
This paper proposes a new method for mapping tissue surface model onto arbitrary target surface whilepreserving the geometrical features of the tissue surface. In our method, firstly, the tissue model is roughly deformed by using Self-organizing deformable model. Since the deformed model may contain folded patches, the folded patches are removed. Moreover, FFD and area-preserving mapping are used to map the model onto the target surface while keeping the ratio of the area of each patch to the area of of the whole tissue surface. From several experimental results, we can conclude that the proposed method can map tissue models onto arbitrary target surface without foldovers.. |
| 34. |
ABNORMAL BEHAVIOR DETECTION USING SURVEILLANCE VIDEOS. |
| 35. |
D-16-6 Shape Reconstruction from Stereo Endoscopic Images Using Wide-Range Edge. |
| 36. |
Ryo Kurazume, Kan Yoneda, Shigeo Hirose, Feedforward and feedback dynamic trot gait control for quadruped walking vehicle, Autonomous Robots, 10.1023/A:1014045326702, Vol.12, No.2, pp.157-172, 2002.03, To realize dynamically stable walking for a quadruped walking robot, the combination of the trajectory planning of the body and leg position (feedforward control) and the adaptive control using sensory information (feedback control) is indispensable. In this paper, we propose a new body trajectory, the 3D sway compensation trajectory, for a stable trot gait; we show that this trajectory has a lower energy consumption than the conventional sway trajectory that we have proposed. Then, for the adaptive attitude control method during the 2-leg supporting phase, we consider four methods, that is, a) rotation of body along the diagonal line between supporting feet, b) translation of body along the perpendicular line between supporting feet, c) vertical swing motion of recovering legs, and d) horizontal swing motion of recovering legs; we then describe how we verify the stabilization efficiency of each method through computer simulation, stabilization experimentation, and experimenting in walking on rough terrain using the quadruped walking robot, TITAN-VIII.. |
| 37. |
Geometric and Photometric Integration System for Large Objects
In this paper, we consider the geometric and photometric modeling of large-scale and intricately shaped objects, such as cultural heritage objects. When modeling a large-scale and intricately shaped object, a huge amount of data is required to model the object. We would like to propose two approaches to handling this amount of data : the parallel processing of merging range images and the adaptive algorithm of merging range images. First, we developed a parallel computation algorithm using a PC cluster which consists of two components, the distributed allocation of range images to multiple PCs, and the parallel traversal of sub-trees of octree. Second, we constructed a merged model in adaptive resolution according to the geometric and photometric attributes of range images for efficient use of computational resources. Moreover, we propose a novel method to construct a 3D model with an appearance by taking a consensus of the appearance changes of the target object from multiple range images.. |
| 38. |
Fast Level Set Method and Realtime Tracking of Moving Objects in a Sequence of Images
The level set method, introduced by S. Osher and J. A. Sethian, has attracted much attention as a topological-free active contour model. This method utilizes an implicit representation of a contour to be tracked, and is able to handle the topological change of the contour naturally. Various applications based on the level set method have been presented including motion tracking, 3D geometrical modelling, and simulation of crystallization or semiconductor growth. However, the calculation cost of reinitialization and updating of the implicit function is considerably expensive as compared with the cost of conventional active contour models such as "Snakes". In this paper, we propose an efficient calculation algorithm for the level set method named the Fast Level Set Method (FLSM). Advantages of the proposed FLSM are as follows: i) the use of the extension velocity and the high speed construction of the extension velocity field using the Fast Narrow Band Method, ii) the frequent execution of the reinitialization process of the implicit function which requires little calculation cost. The efficiency of the proposed method is verified through computer simulations, and experiments of realtime tracking of moving obiects in video images.. |
| 39. |
Early Recognition and Prediction of Gestures for Proactive Human-Machine Interface
This paper concerns two topics on gesture recognition. The first topic is early recognition for providing the recognition result of a gesture before the gesture is completed. The second topic is motion prediction for guessing the subsequent posture of the person who makes a gesture. Both topics are mutually related and linked to the realization of proactive human-machine interface. For each of those two topics, a simple technique is developed and examined to reveal its limitation. Possible directions to deal with the limitation are also discussed as the future work on those topics.. |
| 40. |
Parallel Simultaneous Alignment of a Large Number of Range Images on Distributed Memory System
This paper describes a method for parallel alignment of multiple range images. Since it is difficult to align a large number of range images simultaneously, we developed a parallel method to accelerate and reduce the memory requirement of the process. Although a general simultaneous alignment algorithm searches correspondences for all pairs of all range images, by rejecting redundant dependencies, our method makes it possible to accelerate computation time and reduce the amount of memory used on each node. The correspondence search is performed independently for each pair of range images. Accordingly, the computations between the pairs are preformed in parallel on multiple processors. All relations between range images are described as a pair node hyper-graph. Then, an optimal pair assignment is computed by partitioning the graph properly. The method was tested on a 16 processor PC cluster, where it demonstrated the high extendibility and the performance improvement in time and memory.. |
| 41. |
Telecommunication via embodied proactive interface
The purpose of this research is the development of a new interface called "proactive interface" for natural telecommunication. Features of the proactive interface are twofolds. The first feature is an embodied device using robot technology. Instead of virtual media, humanoids are used as the interface for presenting gesture of a user to a distant user. The second feature is estimation of user's intention for compensating system delays. A recognition-based gesture prediction scheme can be used for the estimation. A two-way telecommunication system connecting two distant campuses was developed to demonstrate the proactive interface.. |
| 42. |
Telecommunication via embodied proactive interface
The purpose of this research is the development of a new interface called "proactive interface" for natural telecommunication. Features of the proactive interface are twofolds. The first feature is an embodied device using robot technology. Instead of virtual media, humanoids are used as the interface for presenting gesture of a user to a distant user. The second feature is estimation of user's intention for compensating system delays. A recognition-based gesture prediction scheme can be used for the estimation. A two-way telecommunication system connecting two distant campuses was developed to demonstrate the proactive interface.. |
| 43. |
Ryo Kurazume, Tsutomu Hasegawa, A new index of serial-link manipulator performance combining dynamic manipulability and manipulating force ellisoids, IEEE Transactions on Robotics, 10.1109/TRO.2006.878949, Vol.22, No.5, pp.1022-1028, 2006.10, The inertia matching ellipsoid (IME) is proposed as a new index of dynamic performance for serial-link robotic manipulators. The IME integrates the existing dynamic manipulability and manipulating-force ellipsoids to achieve an accurate measure of the dynamic torque-force transmission efficiency between the joint torque and the force applied to a load held by an end-effector. The dynamic manipulability and manipulating-force ellipsoids can both be derived from the IME as limiting forms, with respect to the weight of the load. The effectiveness of the IME is demonstrated numerically through the selection of an optimal leg posture for jumping robots and optimal active stiffness control, and experimentally through application to a pick-up task using a commercial manipulator. The index is also extended theoretically to the case of a manipulator mounted on a free-flying satellite. © 2006 IEEE.. |
| 44. |
Kazuya Matsuo, Kouji Murakami, Tsutomu Hasegawa, Kenji Tahara, Ryo Kurazume, Segmentation method of human manipulation task based on measurement of force imposed by a human hand on a grasped object, 2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS 2009, 10.1109/IROS.2009.5354069, pp.1767-1772, 2009.12, This paper proposes a segmentation method of human manipulation task based on measurement of contact force imposed by a human hand on a grasped object. We define an index measure for segmenting a human manipulation task into primitives. The indices are calculated from the set of the contact forces measured at all the contact points during a manipulation task. Then, we apply the EM algorithm to the set of the indices in order to segment the manipulation task into primitives. These primitives are mapped onto the robotic hand to impose appropriate contact forces on a grasped object. In the experiments, manipulation tasks performed in daily human life have been successfully segmented. © 2009 IEEE.. |
| 45. |
GUARNIERI MICHELE, guarnieri michele, 倉爪 亮, KURAZUME RYO, DEBENEST PAULO CESAR, Debenest Paulo Cesar, 程島 竜一, Hodoshima Ryuichi, 福島 E. 文彦, FUKUSHIMA EDWARDO FUMIHIKO, 広瀬 茂男, HIROSE SHIGEO, HELIOS System: A Team of Tracked Robots for Special Urban Search and Rescue Operations, 10.1109/IROS.2009.5354452, 2009.10. |
| 46. |
Real-Time Nonlinear FEM-Based Simulator for Deforming Volume Model of Soft Organ by Neural Network
本論文では,ニューラルネットワークを用いて,軟性臓器モデルの変形をシミュレートする新たな手法を提案する.提案手法は,基本的なモデルの変形(以後,変形モードと呼ぶ)の組合せに基づいて,モデルの変形を推定する.つまり,変形モードをあらかじめ非線形有限要素法で求め,臓器に加わった外力と,それに対応する変形モードの関係をニューラルネットワークで学習する.学習したニューラルネットワークは,非線形有限要素解析によりモデルの振舞いを推定することを模倣する.実験結果より,提案手法は,非線形有限要素解析とほぼ同程度の精度を保ちつつ,計算コストを大幅に削減することができた.. |
| 47. |
Articulated Structure from Motion using Subspace Separation
This paper considers the problem of articulated structure from motion where the target object is made of articulated rigid bodies and all the feature points are not always observed simultaneously. We deal with this problem by subspace separation method and it is quite important to use feature points whose accuracy in position and separation is high. In this study, we propose a method that rejects wrong feature points as much as possible so as to estimate structure and motion stably with the remaining good feature points.. |
| 48. |
Articulated structure from motion using subspace separation. |
| 49. |
Articulated structure from motion using subspace separation. |
| 50. |
Articulated structure from motion using subspace separation. |
| 51. |
Person identification using affine moment invariants
Gait is one of biometrics which offer the possibility to identify people at a distance, without any interaction or co-operation from the subject, compared with other kind of biometrics, such as fingerprints and iris. To identify people by gait, gait features based on a model of the human body or one's appearance are extracted from captured image sequences of walking people. Various methods have been proposed so far, and a Fourier transform-based method enabled to identify people with high correct classification rate. However, the dimension number of features rises in parallel with the increasing of the image resolution. In this paper, we introduce a resolution independent method using affine moment invariants, which enables to identify people with high correct classification using small dimension number of features. We applied the proposed method to a gait database SOTON, and used the leave-one-out cross validation technique to estimate the correct classification rate of 98 % with twentieth part of dimension number of features of the conventional method.. |
| 52. |
Oscar Martinez Mozos, Ryo Kurazume, Tsutomu Hasegawa, Multi-part people detection using 2D range data, International Journal of Social Robotics, 10.1007/s12369-009-0041-3, Vol.2, No.1, pp.31-40, 2010.03, People detection is a key capacity for robotics systems that have to interact with humans. This paper addresses the problem of detecting people using multiple layers of 2D laser range scans. Each layer contains a classifier able to detect a particular body part such as a head, an upper body or a leg. These classifiers are learned using a supervised approach based on Ada Boost. The final person detector is composed of a probabilistic combination of the outputs from the different classifiers. Experimental results with real data demonstrate the effectiveness of our approach to detect persons in indoor environments and its ability to deal with occlusions. © Springer Science & Business Media BV 2009.. |
| 53. |
Indoor place categorization for service robots using camera and depth images
An important capability for service robots working with humans in indoor environments is their ability to categorize the different places where they are located. In this paper we present an approach to categorize different areas in indoor environments using an RGB-D sensor like the Kinect camera. First, RGB images are transformed into grey scale images. Then, grey scale and depth images are transformed into histograms of local features that incorporate neigh-boring relations by applying local binary patterns and also a short version of this pattern. The feature vectors corresponding to grey scale and depth images are combined and categorized into different places using supervised classifiers like for example support vector machines. We apply this method to distinguish five different place categories and obtain high recognition rates.. |
| 54. |
Indoor place categorization for service robots using camera and depth images
An important capability for service robots working with humans in indoor environments is their ability to categorize the different places where they are located. In this paper we present an approach to categorize different areas in indoor environments using an RGB-D sensor like the Kinect camera. First, RGB images are transformed into grey scale images. Then, grey scale and depth images are transformed into histograms of local features that incorporate neigh-boring relations by applying local binary patterns and also a short version of this pattern. The feature vectors corresponding to grey scale and depth images are combined and categorized into different places using supervised classifiers like for example support vector machines. We apply this method to distinguish five different place categories and obtain high recognition rates.. |
| 55. |
Gait identification robust to changes in observation angle
In person identification using gait images, various inherent image features of individuals are extracted from a sequence of gait images taken by a camera. However, for instance, if the subject is close to the camera and the camera captures gait images from the side direction, observation angles between walking direction of the subject and directions of the camera to the subject at all frames are varied during one gait cycle. This unfavorable change induces the decrease of the identification performance. So in this paper, we propose a novel gait identification technique which is robust to changes in observation angle in one gait cycle. The proposed technique utilizes a 4D gait database consisting of multiple 3D shape models of walking subjects and adaptive virtual image synthesis. Experiments using the 4D gait database of 21 subjects show that the proposed method is robust to the changes of the observation angles in one walking cycle and achieves higher recognition performance than the case using a fixed observation angle in one gait cycle. Besides, experimental results show the feasibility of identifying people walking on curved trajectories.. |
| 56. |
Gait identification robust to changes in observation angle
In person identification using gait images, various inherent image features of individuals are extracted from a sequence of gait images taken by a camera. However, for instance, if the subject is close to the camera and the camera captures gait images from the side direction, observation angles between walking direction of the subject and directions of the camera to the subject at all frames are varied during one gait cycle. This unfavorable change induces the decrease of the identification performance. So in this paper, we propose a novel gait identification technique which is robust to changes in observation angle in one gait cycle. The proposed technique utilizes a 4D gait database consisting of multiple 3D shape models of walking subjects and adaptive virtual image synthesis. Experiments using the 4D gait database of 21 subjects show that the proposed method is robust to the changes of the observation angles in one walking cycle and achieves higher recognition performance than the case using a fixed observation angle in one gait cycle. Besides, experimental results show the feasibility of identifying people walking on curved trajectories.. |
| 57. |
Gait identification based on shadows from infrared lights
This paper introduces a novel system for person identification from shadow images of walking person projected by invisible lights, and a shadow database of walking people. In general the correct classification rate of person identification is better when multiple cameras from different viewpoints are used, but most of conventional methods have used one camera, because of (i) easy installation in real environments, since there is no need to synchronize cameras, (ii) reduction of calculation costs. In the proposed system, we obtain the advantages of multiple viewpoints but with a single camera. More specific, we install multiple infrared lights to project shadows of a subject on the ground and a camera with an infrared transmitting filter to the ceiling inside of a building. Shadow areas, which are projections of one's body on the ground by multiple lights, can be considered as body areas captured from different viewpoints, so the proposed system enables to capture multiple body areas from only one camera. We collect a shadow database consisting of 28 people with this system, and we extract features from shadow areas by spherical harmonics, followed by identification of the subject. Experiments using the gait database show the effectiveness of the proposed method.. |
| 58. |
Gait identification based on shadows from infrared lights
This paper introduces a novel system for person identification from shadow images of walking person projected by invisible lights, and a shadow database of walking people. In general the correct classification rate of person identification is better when multiple cameras from different viewpoints are used, but most of conventional methods have used one camera, because of (i) easy installation in real environments, since there is no need to synchronize cameras, (ii) reduction of calculation costs. In the proposed system, we obtain the advantages of multiple viewpoints but with a single camera. More specific, we install multiple infrared lights to project shadows of a subject on the ground and a camera with an infrared transmitting filter to the ceiling inside of a building. Shadow areas, which are projections of one's body on the ground by multiple lights, can be considered as body areas captured from different viewpoints, so the proposed system enables to capture multiple body areas from only one camera. We collect a shadow database consisting of 28 people with this system, and we extract features from shadow areas by spherical harmonics, followed by identification of the subject. Experiments using the gait database show the effectiveness of the proposed method.. |
| 59. |
Laser-based modeling of large-scale environment using multiple mobile robots
We have been proposing a high precision laser-based 3D measurement system by multiple mobile robots. This system is composed of three mobile robots consisting of a parent robot and two child robots. The parent robot is equipped with a 3D laser scanner, attitude sensor and a total station, and the child robots are equipped with corner cubes. The parent robot moves and stops repeatedly, and measures the 3D shape using the equipped laser scanner at several positions. Meanwhile, the child robots also move and stop repeatedly, and act as mobile landmarks for the positioning of the parent robot. This paper presents some improvements of the proposed system by replacing and installing several devices to make the positioning accuracy higher. The experimental results show the system achieves quite high accuracy of the 0.03 % of target's size. The omni-directional sensing robot equipped with four RGB-D cameras and its measurement experiments are also introduced.. |
| 60. |
Laser-based modeling of large-scale environment using multiple mobile robots
We have been proposing a high precision laser-based 3D measurement system by multiple mobile robots. This system is composed of three mobile robots consisting of a parent robot and two child robots. The parent robot is equipped with a 3D laser scanner, attitude sensor and a total station, and the child robots are equipped with corner cubes. The parent robot moves and stops repeatedly, and measures the 3D shape using the equipped laser scanner at several positions. Meanwhile, the child robots also move and stop repeatedly, and act as mobile landmarks for the positioning of the parent robot. This paper presents some improvements of the proposed system by replacing and installing several devices to make the positioning accuracy higher. The experimental results show the system achieves quite high accuracy of the 0.03 % of target's size. The omni-directional sensing robot equipped with four RGB-D cameras and its measurement experiments are also introduced.. |
| 61. |
Gait identification robust to changes in walking direction by 4D gait database
In person identification using gait images, various inherent image features of individuals are extracted from a sequence of gait images. However, for instance, if a subject's observation angle changes compared with those in the database, the correct classification rate gets low. To deal with this problem, we constructed a 4D gait database consisting of multiple 3D shape models of walking subjects, and introduce a method robustly against walking direction changes. In this method, firstly we reconstruct 3D models of subjects from gait images taken by multiple cameras, and then synthesize virtual images of 3D models from multiple arbitrary virtual viewpoints and build a database from gait features extracted from virtual images. In the identification phase, the person is identified by matching the gait features of the subject and those from all virtual viewpoints in the database. However, the calculation cost is expensive due to full search, and the subject is wrongly estimated due to wrong estimation of walking direction. So in this paper, to achieve the reduction of calculation time and high correct classification rate, we introduce a method which estimates the walking direction using Frieze Patterns firstly and then identify the person using features from the estimated virtual viewpoint. Experiments using the 4D gait database show the effectiveness of the proposed method.. |
| 62. |
Area- and Angle-preserving Projection of Tissue Surface Model onto Target Surface
This paper proposes a new method for mapping a tissue surface model onto a target surface while preserving geometrical features of the tissue surface. The proposed method is based on Self-organizing Deformable Model (SDM) whose shape is deformed by using a competitive learning and an energy minimization approach. In our method, several landmarks selected from the tissue model can be mapped onto their target positions of the target surface. In addition, our method preserves the area and angle of each patch included in the original model before and after mapping. Using the mapping result, some models with different numbers of vertices are represented with a unified model structure. From several experimental results, we can conclude that the proposed method can resample equally the original model while controlling mapping positions of landmarks.. |
| 63. |
Hexahedral Finite Element Modeling of Human Tissue by Using Growing Self-Organizing Deformable Model
Finite Element Method (FEM) is one of the techniques for deformation simulations of human tissue. The basic concept of FEM is the discretization of an object into FE-model that consists of simple geometry called elements. Although The most common type of elements are tetrahedral elements, Hexahedral elements have an advantage of analysis accuracy. However, automatic generation method of hexahedral elements for complex shape have not been established. This paper proposes a method that generate generate hexahedral finite element model for complex shape using Growing Self-Organizing Deformable Model (GSDM). By several simulations using our proposed model, we confirmed that deformation analysis is performed with higher accuracy and in a shorter time.. |
| 64. |
A Method for Mapping Tissue Surface Model onto Target Surface Based on Self-Organizing Deformable Model Preserving Geometrical Features
複雑な形状をもつ人体組織同士を比較する際,それぞれのメッシュモデルを,形状が単純な目標曲面にいったん写像して,写像先で差異を比較する手法がある.このとき,対象組織で共通の解剖学的特徴が,目標曲面上で同じ位置にあると,他の部位でも対応付けが容易になる.また,組織形状に近い曲面を目標曲面として選ぶことで,写像が単純で直感的になり,解析しやすくなる.しかし,従来の写像法では,写像先を直接的に制御できず,また,従来法の目標曲面は平面や球面のみであり,形状を自由に設定するのは困難である.そこで,本論文では,特に脳表モデルに対し,モデルの写像先を制御しながら,脳表に適した形状の目標曲面へ写像する新たな手法を提案する.まず,自己組織化可変モデル変形法を用いて,モデルを目標曲面上へ写像する.この変形法を用いることで,写像の直接的制御や,モデルと同一位相をもつ形状の目標曲面が使用可能となる.この際,隣接していないモデルの頂点が,目標曲面上で同じ位置に写像されている折り畳みが生じている場合があり,この折り畳みを除去する.次に,モデルの幾何情報の一つである,モデルの表面積に対する各パッチの面積比を写像前後で保存しつつ,目標曲面に脳表を写像する.6個の脳表モデルを用いた実験を行い,提案手法は,特徴領域を特定の位置に写像しつつ,脳表モデルを目標曲面へ滑らかに写像できることを確認した.. |
| 65. |
Low-altitude Tracking Method Using Particle Filter with Radars in a Multipath Environment
It is well known that the problem of multipath propagation arises in the tracking low-altitude targets with a monopulse radar and causes large bias error in the estimated altitude of the target. Since the bias error caused by multipath propagation depends on a large number of parameters such as the frequency of the radar waveform, the actual target altitude, and range, it is difficult to estimate the bias errors. In this paper, we propose an altitude estimation method using particle filter and multipath propagation model. The performance of the proposed method is verified through computer simulations.. |
