A Research For Key Technology Of High Performance Medical Imaging Service System Based On P2P Network

PACS System that based on DICOM standards has get great success in the past few years,which brought great convenience to medical imaging examination and diagnosis.It not only improves the cost-effectiveness of the hospital,reduces the expenditure for patients,but also helps to develop the medical imaging education and research.With the popularity of DICOM standards,functions and performance of original PACS systems have been unable to meet the people's demands.It was proposed to develop the PACS system and establish a medical imaging service system,which can provide some basic services in the middleware way,including medical imaging storage,browsing,searching,consultation and 3D reconstruction and so on.And middle wares of the system can be continuously increasing,so that it can satisfy the dynamic changes of needs in the community.However,in the progress of building the medical imaging service system,there are still many technical problems,and the most critical twos are the network transmitting speed and the computing speed:1,the network transmitting speed.As the examination equipment improvements,the medical imaging examination data is increasing,in particular the multi-slice CT there can be up to GB level of the image examination data for each case.Such a large volume of data will brought tremendous pressure to the storage servers and transmission network,and will seriously affect the network transmission speed and efficiency of the work of doctors;2,in the computing speed,three-dimensional CT image reconstruction, medical image search based on contents,medical imaging computer-aided diagnostic,and many other aspects of medical imaging applications,all of them exist a large number of computing volume,and some even cannot run on a ordinary computer,which greatly limits the expansion of the medical imaging service system.Problems of transmitting speed and computing speed are always the basic contents of computer research,over the years,there have been many international solutions.On the software side,people through research data compression algorithm, reducing data traffic,thereby improving transmitting speed;people also through study the image computing algorithms,reducing the amount of calculation to improve computing speed.On the hardware side,people through improve network equipment,and improve network bandwidth,thus enhancing transmission speed; people also improve computer performance by updating equipments,thereby to enhance the computing speed.It is clearly that these methods have been resolved most of the problems in reality,but in the field of medical imaging,there are still many limitations.On the software side,most of the algorithms that increase speed are usually at the cost of reducing some accuracy of image,but it is not allowed by the medical imaging.On the hardware side,on one hand,both of the upgrading of network equipment and computer equipment,will increase people's expenditure,in particular investments for high-speed network equipment and high-performance computer equipment will bring enormous economic burden.On the other hand,for the reason of that in the medical imaging service systems,either transmission and calculation is not always in a peak period,and not needed by all the places,therefore comprehensive upgrading equipment seems to be a waste of hardware resources. Therefore,relying solely on improving software algorithms and upgrading hardware equipments system can not solve the problems that exist in the building.In recent years,peer-to-peer(P2P) technology has been rapidly developed.As it can reasonably schedule the idle resources of the network,P2P has been well used in many fields of computer,such as distributed storage and sharing,search engines, work-cooperation and live video,and so on.Problems of building medical imaging service system are mostly similar to which of ordinary computers.As the same of internet,there are also many idle resources in hospital networks.Therefore,the author thought to use P2P technology to solve the transmitting problem and the computing problem.In this paper,the author made the following main aspects:1,Build up a medical imaging peer-to-peer network,give the solution for how to manage the peer nodes and how to make direct communication between peers, and provide a basic platform to other medical imaging applications that use P2P.2,Propose a DICOM image transmitting method based on P2P network,which use the distributed redundant storage technology and the P2P transmitting technology.And experiments show that this method can improve image transmitting speed,while reducing the burden of image storage server.To a certain extent,the author's method can solve the transmitting speed problem of building the medical imaging services system.3,Propose a content-based image search method that use P2P network.We separate a big image search task to some small subtask,and send the subtask to peers in the network which have idle CPU resources,in this way,we can union a large number of computers to finish the big search task.Experiments show that the author's method can improve the image search speed,and it provide a new line of thought to solve computing problems of medical imaging.