Catálogo de publicaciones - libros

This work investigates effects of vibrotactile stimuli and auditory feedback in perceiving roughness. One of objectives in the research is to derive functional correlations between perceived roughness and principal elements of vibrotactile stimuli. A tactile display device having a 5x6 pin array has been developed as the test bed for the quantitative experiment. Sandpapers were used as specimens of perceived roughness. The experimental results show that there is a tight correlation between perceived roughness and vibrotactile stimuli and their functional relations have been derived. The second objective of this research is to analyze the relation between perceived roughness and auditory feedback. Real sound samples recorded during rubbing surface of sandpaper specimens have been adopted as auditory feedback elements. The experimental results showed human could estimate surface roughness from auditory information if and only if he/she recognized the source of sounds. The results in this papers show possibilities to fortify representing roughness by using a tactile-auditory display in a tangible environment beyond conventional approach changing sizes of grits or grooves on a surface.

China is running to the aging society. In China, the number of the elder’s families of empty nest presents an ascendant trend. It has caused various kinds of demands of the service for spiritual consolation. This article introduces how to distinguish different emotion form the sound sample by affective computing and brings forward the concept of spiritual consolation object. Elder can communicate with this object when they need spiritual consolation. With this, we can set up an interchange environment considering people first, which is intelligent and understandable between children and the elder.

High-fidelity simulations for dynamics and movement control of the human arm is a complex and challenging problem. To realize such high-quality response in real time using today’s computing power is simply not possible. In this paper we have presented a simplified, neural network-based dynamics model of a digital human arm for simulating its reaching movements. The arm is modeled as an assemblage of 2 rigid bodies connected by joints and muscles. A Hill-type model is used to capture the muscle forces. Activation dynamics via a low-pass filtering operation is applied to trigger musculotendon contractions. Since we are interested in real-time macro response, we have replaced the forward dynamics module with a recurrent neural network approach. The results of our simulation show that the model produces fairly realistic reaching motion of the arm and it is able to do so in real-time.

The ability to recognize emotion is one of the hallmarks of emotional intelligence. This paper proposed to recognize emotion using physiological signals obtained from multiple subjects without much discomfort from the body surface. Film clips were used to elicit target emotions and an emotion elicitation protocol, verified to be effective in the preliminary study, was provided. Four physiological signals, electrocardiogram (ECG), skin temperature (SKT), skin conductance (SC) and respiration were selected to extract 22 features for recognition. We collected a set of data from 60 female undergraduates when experiencing the target emotion. Canonical correlation analysis was adopted as a pattern classifier, and correct-classification ratio is 85.3%. The research indicated the feasibility of user-independent emotion recognition using physiological signals. But before emotion interpretation can occur at the level of human abilities, there still remains much work to be done.

In this paper, a new method to build an adaptable body shape considering the motion parameters. The proposed method constructs MPEG conforming scalable virtual human by simplifying the body parts which have similar motion parameters while preserving details with characteristic motion parameters. Characteristic motion is classified by its high deformation over time which is extracted from skinning parameter distribution. The proposed method is adaptable with in MPEG framework and is designed to minimize load of adaptation server which keeping compactness as much as possible. It is also targeted to cover devices with small computational power such as PDAs. Example of body simplification is given to show the effectiveness of the proposed method.

Surgical training systems based on virtual reality simulation techniques offer a cost-effective and efficient alternative to traditional training methods. This paper describes a virtual reality system for training robot assisted neurosurgery. The training system undertakes the task of building the 3D environment of the virtual operation room and the surgical robot. The 3D model of the patient’s brain is reconstructed based on the 2D images. After the virtual robot move to the entrance pose that is planned by the surgeon, the user can control the virtual needle to insert into the virtual brain to verify his surgical plan. 3-D display device and home-made human-computer interactive devices increase the immersion and telepresence of this system and establish a robot assisted surgical training demo system oriented clinical application.

In order to compare the effect of VGLS (Virtual Geometry Learning System) with that of the traditional teaching method in improving students’ spatial ability, we selected two classes (total 106 students) in Jingqiao middle school as experimental class and control class, and pre-tested their basic level of spatial ability. The VGLS was applied in the experimental class for four weeks, while the traditional teaching method was applied in the control class. Some conclusions can be drawn from the result of the post-test: Compared with traditional teaching method, VGLS approach is more significant effective in the improving student’s mental folding, unfolding as well as rotation ability of students, without obvious superiority in improving student’s pattern recognition; VGLS is significant effective in improving females’ mental rotating ability and males’ mental folding and unfolding ability.

In this paper, we propose an example-driven approach to transfer deformation from source meshes to a target character based on sketch. Our approach consists of the following steps: (a) a sketch based mapping between parts of the sources and corresponding portions of the target; (b) deforming the portions of the target based on the deformations of the corresponding parts of the sources by means of differential mean value coordinates; (c) stitching the deformed portions of the target together smoothly by minimization to form a key frame of the target animation. We provide new research contributions on all these topics, and integrate them into our newly developed prototype animating system. Our approach is general and does not require the sources and target to share the same number of vertices or triangles, or to have matching connectivity. Our approach is intuitive and is able to produce highly authentic 3D animations. We demonstrate our approach by constructing several animating examples produced from our prototype system.

Extracting the geometric knowledge of space welding seam from virtual models, which representing the real object to be welded, is one of the fundamental issues for virtual teach-in of robotic welding system. An interactive approach to deal with this issue is described in this paper. This interactive approach is based on accurate real-time collision detection and taking advantage of virtual force reflection, and it is particularly suitable for determining complex space welding seam. The proposed approach can be used as a handy and accurate tool to find complex space welding seam for subsequent automatic generation of task planning and path planning of robotic welding system. We demonstrate the effectiveness of this approach with implemented examples.

In the area of Content-Based Image Retrieval (CBIR), feature extraction and matching are the key techniques for image retrieval. In order to overcome the drawbacks of commonly used color features, an approach to extract spatial layout feature is presented which is based on images’ primary color, and a new algorithm is also presented for image features matching by combing the spatial layout feature and the color histogram feature. The experimental results show that the new algorithm can achieve much better performance for images with typical spatial layout feature.