Figure: The MSS hypergraph structure and the intermediate model of the annotated medical report.
Figure: MDS derived from MSS shown in Figure 9.
Figure: MCS derived from MDS shown in Figure 10. The places enclosed by dotted line represent objects processed in the transformation.
A medical doctor Smith is preparing a report for his supervisor Kessler and his group members Wang and Larson. His report C0 contains a text file of a medical record, an audio memo recorded by himself, and a nuclear image of the patient. The text also refers to a confidential report C2 for his supervisor Kessler only. Suppose Smith receives two annotations about the image in his proposal. One annotation is a text while the other annotation C1 is composed of an animation, an audio, and another text where the audio will start to play two seconds after the animation has been played. Figure (a) shows the MSS of report C0 proposal which includes all the different types of nodes and links in the hypergraph structure to specify the static structure of the multimedia objects. The MSS modeled by the hypergraph structure can be transformed into the corresponding MDS according to the MSS-to-MDS algorithm described in Section 3.2. Figure (b) illustrates the intermediate model derived from the MSS. The number after the comma in each node represents the value of the delay attribute of the object. Figure is the MDS constructed from the intermediate model. Suppose that the receiving end does not have audio device and the bandwidth of the communication network is not enough to transmit the image object unless the size of the image is reduced in order to obtain the guarantee form the QOS manager. By applying Algorithm 2, MCS is transformed from MDS by deleting the audio objects and then employing progressive transmission on the image object. The resulting MCS is shown in Figure in which a two-level progressive transmission is applied on the image object, where image.1 is an image of a lower resolution to be transmitted first while image.2 and image are the data to be transmitted progressively in order to fill up the detail of the original nuclear image. Optimization is possible to reduce the complexity of the G-Net of an MDS. Various heuristics can be devised based on properties of the application. Figure shows an optimized version of MDS in this example by deleting places with delay 0 and their output transitions.
Figure: Optimized version of MDS shown in Figure 10.