Towards Ubiquitous Access of ComputerAssisted Surgery Systems

1
Towards Ubiquitous Access of Computer-Assisted
Surgery Systems

• Hui Liu , Hanping Lufei, Weisong Shi, and
  Vipin Chaudhary
• Xidian University, Wayne State University
• Mobile and Internet SysTems Laboratory
• Department of Computer Science
• Wayne State University
• http//mist.cs.wayne.edu

The 28th IEEE Engineering in Medicine and Biology
Society (EMBC), 2006
2
Outline

• Motivations
• Background
• Design and Implementation
• Results and Evaluations
• Conclusions and Future work

3
Motivations
4
Outline

• Motivations
• Background
• Design and Implementation
• Results and Evaluations
• Conclusions and Future work

5
Background

• Project introduction
• Supported by Michigan Life Science Corridor
  (MLSC).
• Project objective Develop infrastructure to
  advance research in CAS in Michigan and
  elsewhere.
• Preliminary work CASMIL
• CASMIL is a comprehensive Image-guided
  Neurosurgery System with extensive novel
  features.
• CASMIL has been successfully developed at Wayne
  State University in the last three years.
• CASMIL is still a stand-alone CAS system.

6
Outline

• Motivations
• Background
• Design and Implementation
• Results and Evaluations
• Conclusions and Future work

7
Overview

• Application scenario
• Architecture
• Server
• Proxy
• Client
• Strengths
• Provide CAS functionalities on various user
  devices
• Provide QoS by introducing optimized mechanisms

8
Server Design

• Traditional CASMIL functionalities
• Rigid and non-rigid body co-registration
• Automated 3D segmentation
• Brain shift predictor
• Knowledge based query tools
• Intelligent planning
• Augmented reality
• High Performance Computing

• Extension CASMIL functionalities
• Asynchronous socket is used to strengthen
  servers scalability.
• Adaptive compression is introduced to reduce
  the user latency.

9
Proxy Design-I

• The proxy is introduced to improve latency and
  scalability of UbiCAS.
• Three optimization components are introduced.

10
Proxy Design-II

• Caching and Prefetching
• Least Recently Used (LRU) caching
• User-transparent prefetching
• Concurrency Support
• Pre-connection
• Pre-spawn
• Dynamic scheduling
• Compression and Deployment
• Tradeoff existed in heterogeneous environment.
• e.g., compressing/decompressing time,
  compression ratio, bandwidth
• Compression algorithm selection and deployment
  affect latency.
• We test Bitmap-Diff, RAR, ZIP and 7Z compression
  algorithms on UbiCAS.

11
Client Design-I

• User-friendly User Interface
• Various functionalities and devices demand
  particular user interfaces (UI).
• For example, Laptop/Desktop vs. Pocket PC.

(b)
(a)
12
Client Design-II

• Functionalities on Laptops or Desktops
• Functionalities on handhelds devices
• remote DICOM images download
• local DICOM image validation, reading and
  displaying
• advanced image processing functions
• GUI interfaces to execute complicated
  neurosurgical procedures on the UbiCAS Server.
• Segmentation
• Registration
• Planning

13
Client Design-III

• Cross-Platform Compatibility
• Java technology is used on client-side for its
  platform-independent nature.
• By selecting the corresponding Java Virtual
  Machine, UbiCAS Client can run on various
  devices.
• Pseudo-Streaming Model
• Inspired by streaming technology
• To make the surgeon instantly observe the
  continuous slices one by one without being aware
  of the downloading action executed simultaneously
  at the back-end.
• Implementation based on the two-thread model on
  client-side

14
Security and Privacy

• From the privacys perspective
• role-based access control model is adopted to
  guarantee different doctors have different views
  of the same data set.
• From the data integritys perspective
• information should not be altered during
  transmission.
• we use the opaque tokens generated by the server
  to handle these security issues.

15
Outline

• Motivations
• Background
• Design and Implementation
• Results and Evaluations
• Conclusions and Future work

16
Results and Evaluations

• So far, we have developed and tested UbiCAS on
  Pocket-PC, Laptop and Desktop in a controlled
  environment.
• We build a test-bed by using NISTNet network
  emulator.
• We show the segmentation user latency in two
  kinds of network connections, DSL and WLAN, on a
  Pocket PC.

(a) Experimental test-bed.
(b) Segmentation latency.
17
Outline

• Motivations
• Background
• Design and Implementation
• Results and Evaluations
• Conclusions and Future work

18
Conclusions and Future Work

• Conclusions
• We propose a novel mobile CAS system, UbiCAS,
  which provides the convenience, safety and
  security for doctors at remote locations to
  access and plan the surgery, share patient
  information in real-time before, during and after
  the surgery.
• Besides functionality implementations on
  heterogeneours devices, optimization mechanisms
  are introduced to guarantee system QoS.
• Future Work
• Deployment of entire functionality of CASMIL and
  testing the correctness of each function and UI
  user-friendliness on other devices
• Deployment and testing of the system in real
  scenarios.

19

• Questions ?