【文章內(nèi)容簡(jiǎn)介】 s as belonging to a single gesture event. To facilitate robust event recognition from camera tracking data, a publishersubscriber design pattern was implemented to distribute event generation responsibilities over multiple functions, allowing each function to have a very limited scope. Each subscriber checks if the raw input from the puter vision ponents of the system meet its specific criteria to generate an event. To allow each application to control the events the system recognizes and to have multiple modes of operation within an application, subscribers are dynamically added and removed at runtime by the client application. Client applications need only handle events. The functions that recognize events can be treated as a black box. Data from both the overhead skin tracking and infrared touch tracking is handled simultaneously by data listeners, allowing for events to incorporate both data sources into their event detection criteria. Figure A2: System Dataflow。 Red in the sychronization phase indicates erroneous points removed.Tracking hands creates two important functionalities, the first is the ability to associate a touch with a user. Multitouch interfaces inherently accept interaction from multiple users, however, it has been impossible so far to discriminate between users in the discreet manner of puter vision skin tracking. User identities are maintained by the overhead tracking information and are not dependent upon an individual39。s skin tone. By having one user standing on each side of the table, each touch can be associated with a corresponding user with very little risk for error. With the added information of which hand created a touch, role specific functions can be given to different users for increased collaborative power. Users should avoid overlapping hands to prevent occlusions in the overhead camera39。s line of site, which may interfere with tracking. The other main benefit is that because a hand is continuously tracked, multiple touches created by the same hand can be treated as part of the same event. By doing this, touches no longer need to be treated as singular events, but can be given a history and associated with previous touches. Furthermore users can now interact with objects on the table without physical contact. One such usage was to have a checker float in front of the users hand after touching it in a simulation of the checkers board game.Additionally, hand location data can be used to increase the touch sensitivity without increasing error. The grayscale image is converted to binary by paring each pixel with a variable threshold value. A lower threshold creates more false positives due to random noise in the image. By tracking hand locations the system automatically ignores erroneous touch data by removing touch locations with large distances from any hand. To allow the system to simultaneously track multiple hands, touches, and run putationally intensive programs, we adopted a distributed architecture so that each camera is connected to a separate puter, each of which municates over a standard network connection. This allows the system to be responsive to user actions with minimal delay while maintaining a high frame rate in the application using only average hardware.Both cameras are low cost webcams running at a 640 x 480 resolution with frame rates of approximately 25 fps for touch tracking and 16 fps for overhead hand tracking under regular usage. The representation of touches as single points based on their center and hands as finger location points and table edge points offers enough precision for actions such as button presses while minimizing memory and network usage.Figure A3: Red reflects touch points determined to be noise。 Blue shows user A touch。 Green shows user B touch。 White reflects detected skin。 Yellow circles indicate the acceptable spatial region for a touch to occur5. Conclusions and Future Work Augmenting a FTIR multitouch table with puter vision skin tracking improves collaboration, expands the methods of interaction and increases touch tracking robustness. The overhead camera is able to discreetly track multiple users by establishing identities for each individual when they make their initial actions and tracking them for the remainder of their usage. Maintaining user identities has two important aspects: preventing users from inadvertently interfering with each other and delegating role specific functions to different users. When two or more people are using a FTIR interface, touches can be confused by the system and unintentional actions can be performed, but this is overe when the system can differentiate touches between users. In many collaborative projects it is mon for people to have different specialties and assignments, efficiency is increased by allowing different users to interact with the same object or region and have different actions performed. The system is also able to recognize gestures that occur above the table, therefore interactions are not relegated to physical contact with the table. Objects on the table can be picked up and moved in a very natural manner because the object follows the hand that selected it. Furthermore, a whole new assortment of natural gestures can be integrated, reducing the amount of cumbersome gestures a user needs to learn.This technology was originally developed for, and is showcased well, within the dynamic environment of m