부산대학교 도서관

Pen stroke data collected on electronic whiteboards is typically collected together and grouped into static pages of information. However to be able to use electronic whiteboards for virtual collaborations in a network environment, the inherent dynamic time based aspect of the stroke media needs to be taken into account. Furthermore given that a full whiteboard contains a rather large amount of data methods need to be devised to efficiently transmit such data in a progressive fashion. In this paper the media properties for stroke data are analyzed and a form is devised where it can be segmented into hierarchical representations of visual detail and made ready for progressive streaming. Strokes are first segmented into non-intersecting substrokes, and then further subsampled in such a fashion as to minimize the error between the original stroke and the subsampled stroke. The error measure used is the area between the original stroke and the subsampled stroke. These substrokes form a progressive more detailed approximation of the original stroke and can be streamed in that order allowing progressively finer display of whiteboard materials.