SNAPPY Libraries for Surgical Navigation

Presentation from RSLondonSouthEast, 2020

Abstract

Surgical navigation combines medical imaging, medical image computing, registration, tracking, visuali
sation, and graphical user interfaces. Researchers working in this area rely on open source platforms
that combine functionality in these areas to deliver prototype clinical applications. In this talk I w
ill discuss some of these platforms, and our experiences in using them for clinical application develo
pment. I will then introduce the SNAPPY libraries which form a more loosely coupled set of stand-alone
 open source Python libraries. The SNAPPY libraries are intended to support rapid prototyping of clini
cal application, faster translation from prototype to approved clinical software, and better dissemina
tion of research.   

The established platforms in this field are both large and often difficult to install or compile. It i
s becoming increasingly hard to recruit researchers with the skills in C++ development to build clinic
al software using these platforms. SNAPPY is intended to address these problems by providing smaller l
ibraries, written predominantly in Python. Development is accompanied by testing, documentation and qu
ality control to ensure that not only can novel applications be prototyped quickly, but that prototype
s can be developed into clinical applications in under two years.

SNAPPY currently consists of 13 libraries providing functionality for visualisation and augmented real
ity in surgery, together with hardware interface with video, tracking, and ultrasound sources. The lib
raries are stand-alone, open source, and provide Python interfaces. This design approach enables fast
development of robust applications with minimal and well managed dependencies.