Application Of Distributed Parallel-Computing On The Re-Sampling Of Digitized Human

The research of digitized human is a comprehensive research of medicine, information technology and virtual reality technology. The information of human body is studied, then the structure and function can be digitized and visualized, finally the function of human body can be simulated accurately.When the model of digitized human is built, how to deal with the huge data becomes a big problem with the improvement of technology of transverse sections, so it is importance to find out the new technology to deal with this problem.Firstly, in this paper the application software of re-sampling of digitized human was analyzed, and the research found that the application software could complete the re-sampling only on a single computer. The time was short when the re-sampling was dealt on the giant computer, but the time was too long to wait while the re-sampling was dealt on the personal computer, so it was difficult to apply the data of digitized human to the teaching and clinic.In order to resolve this problem, the theory and technique of re-sampling of digitized human were studied in this paper. The data of digitized human was re-sampled with the technique of GDI+. When the data was re-sampled, the technique of pipelining was used. The disk I/O and image operation were dealt with parallel-computing, and the speed of re-sampling was increased to a certain extent, but the speed was still not satisfaction and the technique needed to be improved.Secondly, in order to resolve the problem of serial-computing, the distributed parallel-computing was studied in this paper, and it was more quickly to re-sample the data of digitized human by parallel-computing. In this paper the data of digitized human was re-sampled by parallel-computing with lots of personal computers in the campus. The result of experimentation showed that the time of re-sampling decreased greatly and the speed increased with the distributed parallel-computing for re-sampling transverse sectional anatomical data from digitized human, but the serial processing and communication restrict further parallel processing while the number of clients increased.Lots of clients and computer network maybe met problems sometimes in the distributed parallel-computing, so the distributed parallel-computing had to resolve these problems. The high retractility and expansibility of the distributed parallel-computing was explored in this paper, and the system could fit in with the change of clients. The technique of dynamic load-balance was explored in this paper, but the efficiency of arithmetic needed to be improved.As was showed in the experimentation, the research in this paper could fit in with the change of clients and resources of calculation, and the time of re-sampling decreased greatly and the speed increased. This research can promote the three dimensional researches of digitized human and has comparatively high application value.