A Cloud-based System For Medical Image Processing

As the medical imaging is widely used in modern medicine,medical image processing technology has gotten a lot of attention in the past decade or so.At present,medical images are usually processed by the software system that ships with the imaging machine clinically.There are some shortcomings of these kinds of software:1.Usually,the software must run on an expensive graphics workstation the medical images are characterized by larger amount of data and the processing requires powerful computational resources.It brings the drawbacks such as high cost,low efficiency,being restricted by time and place.2.The learning curve of the software is cliffy and since there are operational variations from one vendor to another,it is more possible that there is a special expert to process the images.When the software need be maintained and updated,an engineer need to be sent on site.3.The software is not suitable for personalized medicine because of indivisual differences and variety of medical imaging.In recent years,the concept of Software as a Service(SaaS)has been getting the recongnition as the Cloud Technology develops.Cloud means the resouces on server side including hardware(processing units,storages and so on)and software(developing environments,application tools and so on).With cloud technology,when the local clients send messages and request tasks to the servers,the clusters of computers on the server side will offer the resources and return the results.In this model,the computational resources and the application tools become a kind of public resources to the users who need it such as water and electricity.To the implementation of a cloud-based medical image processing system,this dissertation proposed a cloud-based processing architecture and designed a prototype.The traditional medical image processing systems depend on high-performance graphics workstations,and hardly run on personal computers or mobile device.But within a cloud-based system,the browser on the clinet side is enough for the whole functions because the storage and computation are totally completed on the service side.In this system,data and processing procedures are separated to make full use of the hardware performance and good expansibility and scalability.With Redis,an in-memory data structure store,the system has engineered a rapid data exchange between servers to respond to requests speedily.The 2D/3D visualization of medical images on browser are supported in the system.Interactive operations such as rotation,zoom-in,translation,window,pseudo-color are implemented on browser without downloading the whole series data.As a web application,the system gets almost the same user experience as a local application.In the framework of the proposed cloud-based medical image processing system,this dissertation proposed a cloud service for radiologists and surgeons for the preoperative planning.The cloud-based liver surgery planning system covers diverse processing algorithms like liver segmentation,tumor segmentation,intro-hepatic vessels extraction and modeling.This service has been initially proved in living donor liver transplantation and giant mess type hepatectomy.