Liangbing Feng

Extending touch-less interaction on vision based wearable device

A touch-less interaction technology on vision based wearable device is designed and evaluated. Users interact with the application with dynamic hands/feet gestures in front of the camera. Several proof-of-concept prototypes with eleven dynamic gestures are developed based on the touch-less interaction. At last, a comparing user study evaluation is proposed to demonstrate the usability of the touch-less approach, as well as the impact on users emotion, running on a wearable framework or Google Glass.

A gesture recognition system with retina-V1 model and one-pass dynamic programming☆

Abstract Dynamic Programming (DP) algorithm has been studied from 1940s and successfully applied to pattern recognition fields such as continuous speech recognition, hand writing recognition, gesture recognition and so on. In this paper, we propose a novel hand gesture recognition system which includes three kinds of image processing: skin area segmentation, motion estimation by a retina-V1 model, and a gesture discrimination algorithm of One-Pass Dynamic Programming (One-Pass DP). A HSV - RGB filter is used to extract skin area in the color image, and the simple motion of hand area is estimated in eight directions by a retina-V1 model which is a computational model of primary visual cortex. Then the motions are used to compose 40 basic templates of gestures. In other words, hand gestures are considered as combinations of templates of simple motions, and One-Pass DP is used to recognize the pattern of gestures. Experiments dealt with individual and compound gestures were executed by online processing, and the results confirmed the effectiveness of the proposed system.

An Improved Internal Model of Autonomous Robots by a Psychological Approach

To realize the autonomous exploration and the cooperation behaviors of robots in the unknown environment, an improved internal model to evoke robots actions using a psychological theory of Russell was proposed in our previous work. The improved model is based on an affect-action model proposed by Ide and Nozawa group whose basic principle is to control the movement of robots by the degrees of “pleasure” and “arousal” of one’s own and the observation of others. To overcome the phenomena of “deadlock” and adapt to the complicated environment, “curiosity” factor is introduced into the basic model, and the action function is improved to be dynamically. This paper provides experimental comparison between the conventional model and our improved model with goal-exploration simulations. The results showed that only robots with the improved model moved dynamically and successfully reached at multiple goal areas avoiding local traps and obstacles in the complicated environment.

Autonomic behaviors of swarm robots driven by emotion and curiosity

This paper proposes an improved internal model with emotional and curious factors for autonomous robots. Robots acquire adaptive behaviors in the unknown environment according to make observation of behaviors of others. Cooperative relation among the robots and transition of curiosity to the local environments drive robots to achieve the goal of the environment exploration. Simulations showed the effectiveness of the proposed model with interesting motions of robots.

A Human-Machine Interaction System: A Voice Command Learning System Using PL-G-SOM

This paper proposes a voice command learning system for partner robots acquiring communication ability with instructors. Parameter-less Growing Self-Organizing Map (PL-G-SOM), an intelligent pattern recognition model given by our previous work, is used and computational feeling of robots is also adopted to improve the human-machine interaction system. AIBO robot was used in the experiment and the results of real environment showed the effectiveness of the proposed methods

Intelligent agent construction using the attentive characteristic patterns of chaotic neural networks

Humans learn from incidents in their own life and reflects these in subsequent actions as their own experiences. These experiences are memorized in the brain and recollected when necessary. This research incorporates this type of intelligent information processing mechanism and applies it to an autonomous agent. In the proposed system, the reinforcement Q-learning method is used. Autoassociative chaotic neural networks are also used as mutual associative memory systems. However, an agent cannot retrieve all stored patterns exactly, especially in the case of too many stored patterns and a strong correlation among them. To solve this problem, we propose to use types of attentive parameters and attentive characteristic patterns. The attentive characteristic pattern is part of the stored patterns. When robots concentrate their attention on a specific part of a stored pattern, i.e., the attentive characteristic pattern, whole stored patterns are retrieved easily and completely. Finally, the effectiveness of the proposed method is verified through a simulation applied to plural maze-searching problems.

Plane surface detection and reconstruction using segment-based tensor voting

A Segment-based Tensor Voting (SBTV) algorithm is presented for planar surface detection and reconstruction of man-made objects.Tensor voting is used for obtaining the geometry attribute of the 3D points cloud. The candidate planar patches are obtained the geometry attribute through Tensor voting.We over-segment the scene image into the segment and the candidate 3D planar patch is generated.The SBTV algorithm is used on 3D points cloud sets to identify the co-plane on the candidate patch. A Segment-based Tensor Voting (SBTV) algorithm is presented for planar surface detection and reconstruction of man-made objects. Our work is inspired by piecewise planar stereo reconstruction. During the vital procedure to detect and label the planar surface, the two main contributions are: first, tensor voting is used for obtaining the geometry attribute of the 3D points cloud. The candidate planar patches are generated through scene image segment of low variation of color and intensity. Second, we over-segment the scene image into the segment and the candidate 3D planar patch is generated. The SBTV algorithm is used on 3D points cloud sets to identify the co-plane on the candidate patch. After detecting every planar patch, the geometry architecture of object is obtained. The experiments demonstrate the effectiveness of our proposed approach on either outdoor or indoor datasets.

Construction and Application of Learning Petri Net

Petri nets are excellent networks which have great characteristics of combining a welldefined mathematical theory with a graphical representation of the dynamic behavior of systems. The theoretical aspect of Petri nets allows precise modeling and analysis of system behavior, at the same time, the graphical representation of Petri nets enable visualization of state changes of the modeled system [32]. Therefore, Petri nets are recognized as one of the most adequate and sound tool for description and analysis of concurrent, asynchronous and distributed dynamical system. However, the traditional Petri nets do not have learning capability. Therefore, all the parameters which describe the characteristics of the system need to be set individually and empirically when the dynamic system is modeled. Fuzzy Petri net (FPN) combined Petri nets approach with fuzzy theory is a powerful modeling tool for fuzzy production rules-based knowledge systems. However, it is lack of learning mechanism. That is the significant weakness while modeling uncertain knowledge systems.

A Neuro-fuzzy Network with Reinforcement Learning Algorithms for Swarm Learning

An internal model of autonomous mobile robots (agent) is proposed in this paper. A TSK-type fuzzy net is used as a classifier of environment information, i.e., the state of an agent, and reinforcement learning methods such as Q-learning, sarsa-learning are used to make multiple agents acquire adaptive behaviors. Goal navigated exploration problem was simulated to confirm the effectiveness of the proposed methods, and the results showed that the new learning methods are more efficient than actor-critic method which was proposed by our previous work.

Handwriting character classification using Freeman's olfactory KIII model

KIII model is an olfactory model proposed by W. J. Freeman referring to a physiological structure of mammal’s olfactory system. The KIII model has been applied to kinds of pattern recognition systems, for example, electronic nose, tea classification, etc. However, the dynamics of neurons in the KIII model is given by Hodgkin-Huxley’s second-order differential equation and it consumes a very high computation cost. In this paper, we propose a simplified dynamics of chaotic neuron instead of the Hodgkin-Huxley dynamics at first, and secondly, we propose to use Fourier transformation with high resolution capability to extract features of time series behaviors of internal states of M1 nodes in KIII model instead of the conventional standard deviation method. Furthermore, paying attention to the point that human brain does visual processing as same as olfactory processing in the sense of information processing, a handwriting image recognition problem is treated as a new application field of KIII model. Through the computer simulation of the handwriting character classification, it is shown that the proposed method is useful by the comparison of experiment results with both computation time and recognition accuracy.