论文成果 / Publications
2012
Clustering web pages to facilitate revisitation on mobile devices
Abstract
Due to small screens, inaccuracy of input and other limitations of mobile devices, revisitation of Web pages in mobile browsers takes more time than that in desktop browsers. In this paper, we propose a novel approach to facilitate revisitation. We designed AutoWeb, a system that clusters opened Web pages into different topics based on their contents. Users can quickly find a desired opened Web page by narrowing down the searching scope to a group of Web pages that share the same topic. Clustering accuracy is evaluated to be 92.4% and computing resource consumption was proved to be acceptable. A user study was conducted to explore user experience and how much AutoWeb facilitates revisitation. Results showed that AutoWeb could save up a significant time for revisitation and participants rated the system highly.
How much to share: A Repeated Game Model for Peer-to-Peer Streaming under Service Differentiation Incentive
Abstract
In this paper, we propose a service differentiation incentive for P2P streaming system, according to peers' instant contributions. Also, a repeated game model is designed to analyze how much the peers should contribute in each round under this incentive. Simulations show that satisfying streaming quality is achieved in the Nash Equilibrium state.
A Scalable Distributed Architecture for Intelligent Vision System
Abstract
The complexity of intelligent computer vision systems demands novel system architectures that are capable of integrating various computer vision algorithms into a working system with high scalability. The real-time applications of human-centered computing are based on multiple cameras in current systems, which require a transparent distributed architecture. This paper presents an application-oriented service share model for the generalization of vision processing. Based on the model, a vision system architecture is presented that can readily integrate computer vision processing and make application modules share services and exchange messages transparently. The architecture provides a standard interface for loading various modules and a mechanism for modules to acquire inputs and publish processing results that can be used as inputs by others. Using this architecture, a system can load specific applications without considering the common low-layer data processing. We have implemented a prototype vision system based on the proposed architecture. The latency performance and 3-D track function were tested with the prototype system. The architecture is scalable and open, so it will be useful for supporting the development of an intelligent vision system, as well as a distributed sensor system.
Inertial Body-worn Sensor Data Segmentation by Boosting Threshold-based Detectors
Abstract
Using inertial body-worn sensors, we propose a segmentation approach to detect when a user changes actions. We use Adaboost to combine three threshold-based detectors: force/gravity ratios, peaks of autocorrelation, and local minimums of velocity. Experimenting with the CMU Multi-Modal Activity Database, we find that the first two features are the most important, and our combination approach improves performance with an acceptable level of granularity.
Watching you moving the mouse, i know who you are
Abstract
Previous research on modeling human's pointing behavior focuses on user-independent variables such as target width and distance. In this work-in-progress, we investigate a set of user-dependent variables, which are drawn from cursor trajectory data and may represent an individual user's unique pattern when controlling mouse movement. Using these features, the 8 users in our experiment can be recognized at a promising accuracy as high as 87.5%.
UI Portals: Sharing Arbitrary Regions of User Interfaces on Traditional and Multi-user Interactive Devices
Abstract
This paper introduces UI Portals, a novel approach to help users share their off-the-shelf applications' user interfaces on traditional and multi-user interactive devices among various platforms. Users can choose an application window or select parts of the window to share. In addition to the traditional single-user mouse-and-keyboard interaction, we provide support for simultaneous interactions on large multi-user interactive surfaces, like tabletops and multi-touch vertical surfaces. We describe the concepts and implementation mechanisms of this approach. Furthermore, we implement UI Portals Toolsets (UIPT), a prototype that demonstrates sharing arbitrary regions of user interfaces among multiple platforms without any change to the application source code. In UIPT, we design a windowing tool dedicated to large multi-user interactive surfaces to fully leverage the benefit of simultaneous interaction. Two typical scenarios demonstrate the utility of UIPT and show how UIPT can help users work with their familiar software applications on different displays and platforms.
AutoWeb: automatic classification of mobile web pages for revisitation
Abstract
Revisitation in mobile Web browsers takes more time than that in desktop browsers due to the limitations of mobile phones. In this paper, we propose AutoWeb, a novel approach to speed up revisitation in mobile Web browsing. In AutoWeb, opened Web pages are automatically classified into different groups based on their contents. Users can more quickly revisit an opened Web page by narrowing down search scope into a group of pages that share the same topic. We evaluated the classification accuracy and the accuracy is 92.4%. Three experiments were conducted to investigate revisitation performance in three specific tasks. Results show AutoWeb can save significant time for revisitation by 29.5%, especially for long time Web browsing, and that it improves overall mobile Web revisitation experience. We also compare automatic classification with other revisitation methods.
Digging unintentional displacement for one-handed thumb use on touchscreen-based mobile devices
Abstract
There is usually an unaware screen distance between initial contact and final lift-off when users tap on touchscreen-based mobile devices with their fingers, which may affect users' target selection accuracy, gesture performance, etc. In this paper, we summarize such case as unintentional displacement and give its models under both static and dynamic scenarios. We then conducted two user studies to understand unintentional displacement for the widely-adopted one-handed thumb use on touchscreen-based mobile devices under both scenarios respectively. Our findings shed light on the following four questions: 1) what are the factors that affect unintentional displacement; 2) what is the distance range of the displacement; 3) how is the distance varying over time; 4) how are the unintentional points distributed around the initial contact point. These results not only explain certain touch inaccuracy, but also provide important reference for optimization and future design of UI components, gestures, input techniques, etc.
Fall Detection on Mobile Phones Using Features from a Five-Phase Model
Abstract
The injuries caused by falls are great threats to the elderly people. With the ability of communication and motion sensing, the mobile phone is an ideal platform to detect the occurrence of fall accidents and help the injured person receive first aid. However, the missed detection and false alarm of the monitoring software will cause annoyance to the users in real use. In this paper, we present a novel fall detection technique using features from a five-phase model which describes the state change of the user's motion during the fall. Experiment results validate the effectiveness of the algorithm and show that the features derived from the model as gravity-cross rate and non-primarily maximum and minimum points of the acceleration data are useful to improve the precision of the detection. Moreover, we implement the technique as uCare, an Android application that helps elderly people in fall prevention, detection and first aid seeking.