|
|
 |
 |
 |
 |
 |
Image Processing
Besides database, technique for image processing is also important.
We use SPIDER (Subroutine Package for Image Data Enhancement and Recognition,
a library for image processing archived by the Electrotechnical Laboratory,
Japan) as one of image processing functions. SPIDER contains more than
400 functions useful for routine image processing, such as edge detection,
calculation of moments, and so on. Our framework provides an interface
between SPIDER and VHP data, and lets non-image-processing-specialists
use SPIDER routines with ease.
Internet Compliance
Also, Internet-compliance is necessary so that every kind of operating
system user can access our environment. We have made it possible by developing
a framework, IOF, which converts data into HTML.
Environment for End-User-Computing
Our environment is based on OpenStep,
which is an object-oriented operating system developed by NeXT
(now Apple). Also, the environment
utilizes a framework of MARIOS,
an OpenStep-based application developed at NTT (Nippon Telegraph and Telephone
Corporation) Access Network Systems Laboratories, Japan. Characteristic
of MARIOS is its thorough objected-orinetedness. With MARIOS,
non-programming-specialist can make application dealing with a large database
by drag and drop operations on a graphical user interface. Because MARIOS
is a specific system for database in NTT, we have decided to develop a
generalized version of MARIOS, on which non-programmars can readily make
applications with a graphical user interface. We call this generalized
version of MARIOS as 'Environment for End-User-Computing', and abbreviate
this to 'environment'.
The environment we have made is composed of two components: one is
a sheet on which applications run, and the other is an assembly
of objects derived from frameworks providing compacted bundles of functions.
Based on these frameworks, objects are instantiated by drag-and-drop operations
onto the sheet. An application is a combination of instantiated
objects working on the sheet.
Sheet
On the sheet, various data can be transferred from one object
to another; these objects have been connected by mouse operation and their
connection have been expressed as a link between them. When objects are
activated, the data transfer is initiated and made visible with animations.
LOF
We call a framework for data handling as Library Object Framework (LOF).
Some bundles of LOF are based on Enterprise Object Framework (EOF) developed
by NeXT, which offers objected-oriented handling of database. Main roles
of the bundles of LOF are interfaces between database and other objects.
That is, LOF fetches data from database and hands the data out to objects
in down stream, and vice versa. When LOF fetches/saves data from/to database,
it refers to a model file that describes relationship between object and
database entity. Using this LOF, we are going to integrate Unified
Medical Language System (UMLS) developed by NLM and VHP image data
toward an object-oriented anatomical digital database as a project of the
G7 Information Society, GHAP(Global
Healthcare Application Project) SP8.
ROF
We name a framework for image processing as Research Object Framework
(ROF). Operations ROF can do include Affin transformation, area calculation,
volume rendering, and so on. Also, we have developed an interface between
images and SPIDER routines as described previously.
SPIDER is originally written in Fortran, and rather difficult to use
in conventional programming methods. But the interface we have developed
for SPIDER makes possible to easily call SPIDER routines from the objects.
This means researchers not familiar with Fortran can utilize many useful
routines in SPIDER for image processing.
IOF
We define a framework which connects other objects with the Internet
as IOF (Internet Object Framework). That is, IOF transforms data from the
upper stream into HTML based on an object linked to the IOF object, and
hands them out to an http server.
|
|
|
|
|