Takanobu Tanimoto

Real Time Search for Similar Hand Images from Database for Robotic Hand Control

The authors propose a system for searching the shape of human hands and fingers in real time and with high accuracy, without using any special peripheral equipment such as range sensor, PC cluster, etc., by a method of retrieving similar image quickly with high accuracy from a large volume of image database containing complicated shapes and self-occlusions. In designing the system, we constructed a database in a way to be adaptable even to differences among individuals, and searched CG images of hand similar to unknown hand image, through extraction of characteristics using high-order local autocorrelational patterns, reduction of the amount of characteristics centering on principal component analysis, and prior rearrangement of data corresponding to the amount of characteristics. As a result of experiments, we realized a processing speed of 30 fps or over, and high-accuracy estimation of human hand shape.

Realtime Hand Posture Estimation with Self-Organizing Map for Stable Robot Control

The hand posture estimation system by searching a similar image from a vast database, such as our previous research, may cause the increase of processing time, and prevent realtime controlling of a robot. In this study, the authors proposed a new estimation method of human hand posture by rearranging a large-scale database with the Self-Organizing Map including self-reproduction and self-annihilation, which enables two-step searches of similar image with short period of processing time, within small errors, and without deviation of search time. The experimental results showed that our system exhibited good performance with high accuracy within processing time above 50 fps for each image input with a 2.8 GHz CPU PC.

Real Time Posture Estimation of Human Hand for Robot Hand Interface

The authors propose a hand posture estimation system in real time and with high accuracy, for robot hand control and human interface with hand motions without no sensors attached to the users. This method searches the similar image quickly from a large volume of previously-sorted image database which contains complicated shapes and self-occlusions of the human hand. Because the system doesn¿t need any special peripheral equipment such as range sensor and PC cluster, and works with a note PC using a single high-speed camera, the user can control and make a dexterous robot hand behave as he/she does, and use the system as an information input device by moving his/her hand and fingers.

Copycat hand - robot hand generating imitative behavior

High speed, high accuracy and uniform processing time are essential in the hand posture estimation, to realize a robot hand capable of instantly imitating human actions. Therefore, in the present study, we have developed a method that enables the searching of similar images at high speeds and with high accuracy and the search involves uniform processing time, even in the case where a large-scale database is used. This is achieved by clustering databases having approximately uniform amounts of data using self-organization, including self- multiplication and self-extinction, and by collating the input images with the data in the database by means of the low-order image characteristics, while narrowing the search space in accordance with the past history. By estimating the sequential images of the finger shape by using this method, we successfully realized a process involving a joint angle estimation error within several degrees, a processing time of 150 - 160 fps, and an operating time without dispersion by using a PC having a CPU clock frequency of 2.8 GHz and a memory capacity of 512 MB.

Copycat hand

The goal of authors’ research group is to develop the robot hand system which can learn dexterous motions of the human hands by watching. We have therefore attempted realtime estimation of posture using high-speed search of similar images from database. However, our previous method has a disadvantage that search time increases in proportion to expansion of data though estimating precision gets high. Moreover, the expansion of data may scatter the processing time if the database is not efficiently constituted for high-speed search. The present study therefore classified all the data into classes by using joint angle information of hand and fingers to let self-organizing and adopted twoor multiplestep search utilizing typical characteristic quantity of the classes. In addition, numbers of data affiliated to each class were intended as uniform by adding the algorism of self-reproduction and self-annihilation to each class in self-organization. We propose a high-speed processing above 150 fps with high-precision without deviation of search time.

Real time search for similar hand images from database for robotic hand control

The authors propose a system for searching the shape of human hands and fingers in real time and with high accuracy, without using any special peripheral equipment such as range sensor, PC cluster, etc., by a method of retrieving similar image quickly with high accuracy from a large volume of image database containing complicated shapes and self-occlusions. In designing the system, we constructed a database in a way to be adaptable even to differences among individuals, and searched CG images of hand similar to unknown hand image, through extraction of characteristics using high-order local autocorrelational patterns, reduction of the amount of characteristics centering on principal component analysis, and prior rearrangement of data corresponding to the amount of characteristics. As a result of experiments, we realized a processing speed of 30 fps or over, and high-accuracy estimation of human hand shape.

Mobile robot control using 3D hand pose estimation

We propose a mobile robot control method using 3D hand pose estimation without using sensors or controllers. The hand pose estimation we propose reduces the number of image features per data set, which makes the construction of a large-scale database possible, as well as estimation of the 3D hand poses of unspecified users with individual differences without sacrificing estimation accuracy. The system involves the construction in advance of a large database comprising three elements: hand joint information including the wrist, low-order proportional information on the hand images indicating the rough hand shape, and hand pose data comprised of low-order image features per data set.

Gesture-world technology: 3D hand pose estimation system for unspecified users using a compact high-speed camera

This technology allows people to control devices such as computers, communications devices, household appliances, and robots by means of everyday gestures without using sensors or controllers, which employs the high-speed and high-accuracy computer vision technology capable of estimating the human hand and arm poses captured by a compact high-speed camera.

Real-time hand posture estimation without deviation of search time

The human hand posture estimation system by searching a similar image from a vast database, such as our previous research, may cause the increase of processing time, and prevent real-time controlling of a robot. Moreover, there should be no deviation of search time for stable robot control. In this study, the authors proposed a new estimation method of human hand posture by rearranging a large-scale database with the self-organizing map including self-reproduction and self-annihilation, which enables two-step searches of similar image with short period of processing time, within small errors, and without deviation of search time. The experimental results showed that our the system exhibited good performance with high accuracy within processing time of 10 ms or less for each image input with a 2.8 GHz CPU PC