论文成果 / Publications

It is common to tile normal sized units together to obtain very large tabletops or interactive digital walls. Usually cameras are adopted to capture the touch input on the surface, which demands for fusing several multi-touch input feeds into a single feed in real time. Common approaches of stitching camera frames into aggregate images do not scale to larger number of cameras. We present uStitch Hub to address these challenges, which is specifically applied in the tabletop domain. It is a fusion mediator which accept detected blobs from multiple cameras that do and don't have overlapped field of view, remapping blobs to the dimension of the entire large tabletop, match blobs with real touches on the surface, and concatenate touch trajectories which are performed over two or more units. We conducted laboratory evaluation of uStitch Hub on its stitching success rate and as well as latency. The results prove that uStitch Hub processes fast and provides not only good stitch on touch trajectories but also low latency.

Most existing remote video communication systems are confined to how to accurately deliver adequate information in real time, but lose sight of communicators' interaction demands. Meanwhile, traditional 2D-based video communications cannot make full use of people's 3D information, which shows great potential for video communication to achieve immersive, natural and efficient interactions with 3D technique. To enhance the immersion sense and expand interaction modes in video collaboration, the design goal for ‘immersive' video communication systems was proposed and a novel, avatar-based remote video collaboration system was designed and implemented. Specifically, using Creative Senz3D depth camera, the proposed system extracted people's foreground images through background segmentation as their video avatars, and located the avatars together in a common virtual scenario. Natural and immersive interactions among people and between people and virtual scenes were also well designed, which expanded modes of interaction and collaboration in video communication. Finally, a user study was conducted and the result indicates that the proposed video avatar-based remote video cooperation mode can effectively enhance people's sense of immersion in telecommunication.

We explore improving support for revisitation in documents by automatically generating bookmarks based on users' reading history. After showing that dwell time and number of visits are not appropriate for predicting revisitations in documents, we model the high-level reading task as a sequence of reading blocks and recognize long-distance scrolls as separators between them. A long-distance scroll is defined as a continuous scrolling action which causes the document to be navigated beyond a one-page distance. We propose a new technique, called the Head–Tail (HT) algorithm, to generate bookmarks at the head and the tail of reading blocks, whose validity is quantitatively verified by log data analysis. Two studies were conducted to investigate this HT implicit bookmarking technique. The first is a controlled experiment that compared the HT algorithm to the widely used simple recency algorithm for generating implicit bookmarks, in terms of revisit coverage ability and distance between bookmarks ...

Parallel web browsing describes the behavior where users visit web pages in multiple concurrent threads. Qualitative studies have observed this activity being performed with multiple browser windows or tabs. However, these solutions are not satisfying since a large amount of time is wasted on switch among windows and tabs. In this paper, we propose the multiple-page view to facilitate parallel web browsing. Specifically, we provide users with the experience of visiting multiple web pages in one browser window and tab with extensions of prevalent desktop web browsers. Through user study and survey, we found that 2-4 pages within the window size were preferred for multiple-page view in spite of the diverse screen sizes and resolutions. Analytical results of logs from the user study also showed an improvement of 26.3% in users' efficiency of performing parallel web browsing tasks, compared to traditional browsing with multiple windows or tabs.

Many eyes-free interaction techniques have been proposed for touchscreens, but few researches have studied human's eyes-free pointing ability with mobile phones. In this paper, we investigate the single-handed thumb performance of eyes-free, absolute position control on mobile touch screens. Both 1D and 2D experiments were conducted. We explored the effects of target size and location on eyes-free touch patterns and accuracy. Our findings show that variance of touch points per target will converge as target size decreases. The centroid of touch points per target tends to be offset to the left of target center along horizontal direction, and shift toward screen center along vertical direction. Average accuracy drops from 99.6% of 2*2 layout to 85.0% of 4*4 layout, and average per target varies depending on the location of target. Our findings and design implications provide a foundation for future researches based on eyes-free, absolute position control using thumb on mobile devices.

We present QOOK, an interactive reading system that incorporates the benefits of both physical and digital books to facilitate active reading. QOOK uses a top-projector to create digital contents on a blank paper book. By detecting markers attached to each page, QOOK allows users to flip pages just like they would with a real book. Electronic functions such as keyword searching, highlighting and bookmarking are included to provide users with additional digital assistance. With a Kinect sensor that recognizes touch gestures, QOOK enables people to use these electronic functions directly with their fingers. The combination of the electronic functions of the virtual interface and free-form interaction with the physical book creates a natural reading experience, providing an opportunity for faster navigation between pages and better understanding of the book contents.

In this paper, we present the design of Hero, a suite of learning tools that combine teacher-created extracurricular challenges with in-class motivational tools to help parents become more involved in their child's education, while also engaging students in their own learning. To inform the design, we conducted field studies and interviews involving 7 primary teachers and 15 different families. We analyzed Chinese parenting styles and problems related to parental involvement, and developed three major themes from the data. We then proposed three design goals and created a high-fidelity prototype after several iterations of user testing. A preliminary evaluation showed that teachers, parents, and students could all benefit from the design.

In this paper, we studied how display size affects human pointing performance given the same display field of view when using a mouse device. In total, four display sizes (10.6, 27, 46, 55 inches) and three display field of views (20, 34 and 45) were tested. Our findings show that given the same display field of view, mouse movement time significantly increases as display size increases; but there is no significant effect of display size on pointing accuracy. This research may contribute a new dimension to literature in describing human pointing performance on large displays.