3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf
-
Upload
shantanu-p-eksambekar -
Category
Documents
-
view
1 -
download
0
Transcript of 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf
![Page 1: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/1.jpg)
1
"3Q Factors to Enhance Big Data Medical Image for
Better Diagnosis"
Raflaa Hilmi Hamid, Ahmed Nabeel Ahmed, Hasanain Mohammed Manji
Under the supervision of Dr.Azizah Bt Haji Ahmad
College of Arts and Sciences
University Utara Malaysia
Abstract — Nowadays, images are employed in
several areas of medicine for early diagnosis. In
this sense, the industry provides accurate models
to obtain X-ray, Computerized Tomography (CT),
Magnetic Resonance Imaging (MRI) and others of
high resolution equipments. However, other
images, such as those related to pathological
anatomy present in many situations poor quality,
big quantity of information to be process and non-
quick in enhancement. This complicates the
diagnostic process. This work is focused on the
quality, quantity and quickness enhancement of
this type of images through a system based on
informatics image system combined with
traditional techniques of image processing. The
results show that the proposed methodology can
help medical specialists in the diagnostic of several
pathologies. Considering that the medical image
as a big data issue.
Keywords— super-informatics, medical imaging,
diagnosis, image processing, big Data.
I. Introduction
To analyze the application values of the
super-informatics image technique system (SIITS)
based on picture archive and communication system
(PACS) in improvement of medical imaging (MI)
diagnosis and also image processing techniques [23].
In normal case MI should compare with
multimillion images in one of the informatics system
Vendor Natural Archive (VNA) until this system find
the conformable diagnosis that fit this image, this
process will take a lot of time and effort and also high
cost which made this techniques seem to be slow, unproductive and ineffective [17].
For many years, big data informatics system
developed, but also the MI gets more big and big,
spatially with the urgent need to get early diagnosis
for disease that contact with human life.
The health care organizations spend more and
more every year in order to be advanced by one step
in this field, since early diagnosis for the disease such
as cancer and blood disease is the key of recovery
and survival [6].
An archive is a location containing a collection of
records, documents or other materials of historical
importance. An integral part of picture archive and
communication system (PACS) is archiving. When a
hospital needs to migrate a PACS vendor, the
completed earlier data need to be migrated in the
format of the newly procured PACS. It is both time
and money consuming [23].
PACS was consisted of medical imaging and data
acquisition components and storage and display
subsystems.
Different imaging modalities in modern imaging
system (e.g., X-ray, Ultrasonography (US), Digital
Subtraction Angiography (DSA), Computerized
Tomography (CT), Magnetic Resonance Imaging
(MRI), positive emission tomography (PET)) could
be processed in PACS in the format of Digital
Imaging and Communications in Medicine (DICOM)
imaging.
The PACS was an integrated system, allowing for
efficient electronic distribution and storage of
medical images and access to medical record data.
PACS of different size were widely used in clinical
research stage and diagnostic imaging. With the rapid
development of imaging technology X-ray, US,
DSA, CT, MRI, PET and other modern imaging
devices formed a huge medical imaging system (Big
Data). One of the most important characteristics of
modern imaging system was mass information and
native digital images. As a result, the pattern of
medical imaging education should change a lot of
correspondingly. The foundation of medical imaging
system was high-quality, a large quantity and quick
process of imaging data. It was also well known that
the diagnosing of medical imaging requires
systematic study of a large number of medical images
[1].
At presents, the potentials of PACS for
diagnosing applications were not fully understand by
![Page 2: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/2.jpg)
2
most organization that concern with medical image.
So, a new and effective diagnosing system was
needed in medical imaging techniques. Combing with
PACS, we constructed an imaging informatics system
based on PACS to improve the diagnosis effect of
medical imaging system.
The new concept of Vendor Natural Archive
(VNA) has emerged. A VNA simply decouples the
PACS and workstations at the archival layer. This is
achieved by developing an application engine that
receives, integrates, and transmits the data using the
different syntax of a DICOM format. Transferring the
data belonging to the old PACS to a new one is
performed by a process called migration of data. In
VNA, a number of different data migration
techniques are available to facilitate transfer from the
old PACS to the new one, the choice depending on
the speed of migration and the importance of data.
The techniques include simple DICOM migration,
prefetch-based DICOM migration, medium
migration, and the expensive non-DICOM migration.
―Vendor neutral‖ may not be a suitable term, and
―architecture neutral,‖ ―PACS neutral,‖ ―content
neutral,‖ or ―third-party neutral‖ are probably better
and preferred terms. Notwithstanding this, the VNA
acronym has come to stay in both the medical IT user
terminology and in vendor nomenclature, and
radiologists need to be aware of its impact in PACS
across the globe [17].
On the other hand, other healthcare organization
goes towards develop the image processing
techniques (such as gray level–transaction, image
filtering, Binarization and segmentation).
Image Processing is a technique to enhance raw
images received from CT, MRI, US and other
devices, placed on satellites, space probes and
aircrafts or images taken in normal day-today life for
various applications.
Various techniques have been developed in Image
Processing during the last four to five decades. Most
of the techniques are developed for enhancing images
obtained from unmanned spacecraft's, space probes
and military reconnaissance flights. Image Processing
systems are becoming popular due to easy
availability of powerful personnel computers, large
size memory devices, graphics software etc.
Image Processing is used in various applications such
as:
Remote Sensing
Medical Imaging
Non-destructive Evaluation
Forensic Studies
Textiles
Material Science
Military
Film industry
Document processing
Graphic arts
Printing Industry
The common steps in image processing are image
scanning, storing, enhancing and interpretation.
Image segmentation is the process that subdivides
an image into its constituent parts or objects. The
level to which this subdivision is carried out depends
on the problem being solved, i.e., the segmentation
should stop when the objects of interest in an
application have been isolated e.g., in autonomous
air-to ground target acquisition, suppose our interest
lies in identifying vehicles on a road, the first step is
to segment the road from the image and then to
segment the contents of the road down to potential
vehicles. Image thresholding techniques are used for
image segmentation [10,11].
Even that the two ways (MI processing and
medical informatics system) seems to be predictive,
they still have many weakness point. The Quality,
Quantity, and Quickness (3Q) are the main factors
that we aim to achieve.
We suggest new system for medical images
processing to facilitate the diagnoses of disease in the
better manner. Using both characteristics of these two
ways, step 1: using one or more image processing
technique to enhance our medical image then step 2:
passing the last enhanced one to the informatics
system, that will be easy and quick to match it with
vendor storage system, which can bring back the
diagnosis and certainly the percentage of this
diagnosis will be higher.
The purpose of this study is to raise quality,
quantity and quickness (3Q) in order to highlight the
diagnosis imaging systems in the healthcare and
medical organizations. Also we will try to understand
and discuss both of (image processing technique and
informatics imaging system) briefly, considering MI
as a big data source. Then combine them together in
order to produce new virtual system called ―super-
informatics image technique system‖ (SIITS).
II. Case Study
With the ever-increasing amount of annotated
medical data, large-scale, data-driven methods
provide the promise of bridging the semantic gap
between images and diagnosis. The goal of this paper
is to suggest new technique to enhance informatics
system (such as PACS and VNA) storage technique
in attempting to avoid the high cost of medical
diagnosing and compare with less quantity of image
also increase the quality of all system [17]. we try to
combine both informatics system and image
processing in order to increase quality, quantity and
quickness of imaging data and this will achieve by
![Page 3: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/3.jpg)
3
enhance the medical image with any of processing
image technique, then send this image to informatics
system to be analysis and detect in order to match it
with one of the storage image in this system and get
the diagnosis that will be more effective and more
guaranteed, see figure 3 Super-Informatics Image
Techniques System (SIITS).
III. Overview of image processing and informatics
image technique
One of main purposes of image processing is to
manipulate pixel values for better visibility. For
example, gray-level transformation and image
filtering are typical image processing techniques for
converting an input image into a new image with
better visibility. Another purpose of image processing
is to extract some target objects or regions from an
input image. For example, if we extract all organelles
of a specific kind, we can count them and also
understand their distribution and behavior in a cell [2,
3].
On the other hand, a general purpose of
informatics image technique (such as PACS or VNA)
to classify an image or a target object or a region into
one of types, i.e. classes. Although it is difficult to
achieve the informatics accuracy of human beings,
informatics image technique has already been used in
various applications. Optical character informatics
(OCI) is one of the most classic applications, where
an image of a single character is classified into one of
52 classes.
Table 1 indicates how to select image processing
and informatics image techniques according to our
purpose. All of those techniques can be applicable to
biological image analysis. Note that there is no strict
boundary between image processing and informatics
image technique. Many intelligent image processing
techniques rely on informatics image techniques [4,
15].
It is rather rare to use a single image processing
technique or a single informatics image technique. In
fact, they are often used in a mixed manner in order
to realize a complete system for a specific task (For
example diagnosis), for extracting target organelles
from an image, an image segmentation technique is
first applied to the image and then each segment (i.e.
region) is fed into an informatics image technique
(such as PACS or VNA) for deciding whether the
segment is target or not. Figure 2 shows an examples
for segmentation technique which applied to medical
image first to find specific target, we need to
understand all functions of individual techniques and
useful combinations of the techniques [25].
It is very important to understand the fact that
medical images are often far more difficult to be
processed and recognized than popular (i.e. daily-
life) images, such as character, face, and person
images. In particular, microscopic bio images have
the following difficulties for image processing and
informatics [1].
Figure 2: Combination of multiple image
processing phase's techniques for realizing a
complete system of image processing.
Image acquisition
as digital form of image
Noise removal
step to easier identify target
Image
Segmentation
Recognition of each
segment for finding the
target segment
![Page 4: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/4.jpg)
4
Table 1: Image processing and recognition methods which fit to a specific purpose
Figure 3: Super-informatics image techniques system (SIITS)
Image Processing
Techniques
Capture Medical
Image
VNA System
Medical Image
storage
The Image Diagnosed
![Page 5: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/5.jpg)
5
IV. The concept of Big Data in image processing
and healthcare
Big Data is the future of image processing which
it is part of healthcare scope need to devote time with
big data poised to change the healthcare ecosystem,
organizations and resources to understanding this
phenomenon and realizing the envisioned benefits.
All healthcare constituents (members, payers,
providers, groups, researchers, governments etc.) will
be impacted by big data, which can predict how these
players are likely to behave encourage desirable
behavior and minimize less desirable behavior. These
applications of big data can be tested, refined and
optimized quickly and inexpensively and will
radically change healthcare delivery and research.
Leveraging big data will certainly be part of the
solution to controlling spiraling healthcare costs. We
will try to define big data, explore the opportunities
and challenges it poses for healthcare organizations
to understand how present the hug quantity of
medical images can be as a Big Data issue [26, 27].
A large amount of data becomes "big data"
when it meets five criteria: volume, variety, velocity,
veracity and value, figure 4.
Figure 4: Five Vs‘
Here is a look at more important three V’s:
Healthcare Big Data: Volume
Big data in healthcare means there is a lot of data
— terabytes or even petabytes (1,000 terabytes). This
is perhaps the most immediate challenge of big data
in medical requirements, as it requires scalable
storage and support for complex, distributed queries
across multiple data sources. The challenge is being
able to identify, locate, analyze and aggregate
specific pieces of data in a vast, partially structured
data set [32].
While standard techniques and technologies exist
to deal with volumes of structured data, it becomes a
significant challenge to analyze and process a large
amount of highly variable data and turn it into
actionable information. But this is also where the
potential of big data potential lays, as effective
analytics allow you to make better decisions and
realize opportunities that would not otherwise exist.
Such examples of large-data, their promise and
challenges, have not gone unnoticed. In the US, The
National Science Foundation, the National Institutes
of Health, the Defense Department, the Energy
Department, Homeland Security Department as well
as the U.S. Geological Survey have all made
commitments toward ―big data‖ programs. The
Obama Administration itself has even gotten in on
the act. In response to recommendations from the
President‘s Council of Advisors on Science and
Technology, the White House sponsored a meeting
bringing together a cross-agency committee to lay out
specific actions agencies should take to coordinate
and expand the government‘s investment in ―big
data‖, totaling $200 million in support; we mention it
all as an example of big data feasibility [30, 32].
Healthcare Big Data: Variety
There are three different forms of data in most
large healthcare institutions. Discretely codified
billing and clinical transactions are well suited for
relational data models. Digital capture and
management of diagnostic imaging studies required
the development of specialized data formats,
communication protocols, and storage systems.
While these PACS systems are not typically
recognized as big data, they clearly meet the criteria
we have outlined here.
The third form of data in healthcare consists of
blobs of text, typically generated to document an
encounter or procedure. While stored electronically,
there is very little analysis done on this data today,
because database servers are not able to effectively
query or process these large strings. Natural language
processing has been around since the 1950‘s, but
progress in the field has been much slower than
initially expected. The accuracy and reliability of the
results produced by this technology do not yet meet
the requirements of most clinical analytic use cases.
There is much opportunity for progress in this area,
particularly for clinical research [29].
Healthcare Big Data: Velocity
The speed at which some applications generate
new data can overwhelm a system‘s ability to store
that data. Data can be generated from two sources:
humans, or sensors. We have both sources in
healthcare. With a few exceptions like diagnostic
![Page 6: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/6.jpg)
6
imaging and intensive care monitoring, most of the
data we use in healthcare is entered by people, which
effectively limits the rate at which healthcare
organizations can generate data. Like a hospital,
Facebook‘s data is generated by people [32].
V. Healthcare Analytics and Deeper Insight
Data analytics, wisely used, can create business
value and competitive advantage. Compared with
many other industries, healthcare has been a late
adopter of analytics. Most health systems have lots of
opportunities to improve clinical quality and financial
performance, and analytics are required to identify
and take advantage of those opportunities.
It‘s a long journey for most organizations to
develop of culture of continuous, data-driven
improvement. Can big data help your healthcare
organization along this journey? Hopefully you now
have a framework to help guide your thinking.
Learn why when it comes to predictive analytics,
sometimes big data is a big miss. Or see why
advanced analytics can‘t solve all of health care‘s
problems [28, 31].
VI. New Computer Science Designed with Big
Data in Mind
While it might be tempting to think that once all
the medical images (MI) data has been archived,
indexed, and is ready to go, that all one would need is
to start analyzing it and answers to all our questions
about the all MI data will be revealed. Even
examining the contents of an archive to know what
data is available to be analyzed requires new, cleverly
designed and user friendly software tools and novel
approaches for exploratory inspection [28]. Such
tools are only now beginning to appear and their
further development will be essential for dealing with
existing as well as the expected size of MI data sets.
Once a selection of data worthy of further analysis
has been identified, a new concern is realized – it
becomes clear that many software packages for MI
data analysis are ill-suited toward very large data sets
involving potentially thousands of subjects.
Algorithm optimization is not often considered for
when data sets are small or modest in size but as data
sets grow memory management is an important
factor. New mathematics and informatics approaches
will be needed to more completely model multi-
modal MI data in the context of diagnosis disease,
white matter connectivity, and functional activity.
These will need to work fast, be accurate, and be
interoperable with other tools so that data processing
can be automated as much as possible. Interactive
workflow environments for automated data analysis
will also be critical for ongoing or retrospective
research studies involving complex computations on
large multi-dimensional datasets. Yet, few tools, if
any, now exist which enable the joint analysis of MI
data which would be capable of efficiently obtaining
results while also achieving the requisite degree of
statistical power. Moving forward, software
engineers will need to create brilliant and innovative
ways to tackle the massive amounts of MI data [31].
VII. Image segmentation
Image segmentation is one of the most important
image processing techniques for medical images. Its
purpose is to partition an input image into regions.
Image segmentation is necessary for multiple
purposes; for example, counting objects, measuring
the two-dimensional (or three-dimensional)
distribution of the objects, measuring the shape or
appearance of individual objects, recognizing the
individual objects, localizing objects for tracking,
removing unnecessary regions, etc [33].
It is important to note that image segmentation is
the most difficult task among all image processing
tasks. Even though human beings perform image
segmentation without any difficulties, computers
often suffer from its difficulty. In fact, we have not
had any perfect segmentation method yet even for
human face separation from a picture. Biological
images often have far more difficulties than face
images. This is because target objects in biological
images have ambiguous boundaries and thus are
difficult to be separated from the background and
other objects. Furthermore, all the difficulties listed
in the Introduction (such as low resolution) make
segmentation a difficult task.
Table 2 lists typical image segmentation methods,
which have been developed for general (i.e. non-
biological) images. Those methods are overviewed
below, except for binarization. Again, there is no
perfect segmentation method, see figure 5, especially
for biological images. It will be an important future
work to develop new methods specialized for
biological images [8, 34].
Figure 5: Various types of segmentation
![Page 7: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/7.jpg)
7
Table 2 List of image segmentation methods
VIII. Medical Image Processing in Healthcare
Industry
Medical image processing needs continuous
enhancements in terms of techniques and applications
to help improve quality of services in health care
industry. The techniques used for interpolation,
image registration, compression, medical diagnosis
are to be improved to be abreast with growing
demands in the industry and emerging technologies
pertaining to mobile computing and cloud computing.
From the analysis of the literature it is understood
that the health care domain has got much scope for
further research in the areas of diagnosing life
threatening diseases, usage of remote health
monitoring applications for real time functioning to
alert healthcare employees. The integration of
medical equipment and applications with wearable
devices is also promising area for further research [5,
28].
Growing interest in health care domain has paved
way for innovative approaches for medical diagnosis
and clinical practices. Since health is considered to be
wealth, the healthcare industry has been striving to
use innovative medical procedures and treatment
practices coupled with technologies in computations,
harnessing advances in hardware resources [7, 14].
Precision in disease diagnosis and accuracy in
clinical practices and improvement in state-of-the-art
equipment is the ever-ending necessity in the health
care industry. This has led to various best practices
Name Methodology Merit Demerit
Image
binarization
See Table 1 Appropriate when the target
object is comprised of only
bright pixels (or dark pixels)
Limited applicability (however, note that
several binarization methods can be extended
for multi-level thresholding. For example, by
using two thresholds, an image is partitioned
into bright regions, mid regions, and dark
regions.)
Background
Subtraction
Detect target objects by
removing the background
part
Appropriate when target objects
are distributed over the
background
The background image is necessary.
Especially when the background is not
constant, some dynamic background
estimation is necessary
Watershed
Method
Representing an image as a
three-dimensional surface,
and detecting its ridge lines,
i.e. watershed
Even if gray-level change is not
abrupt, it is possible to detect
its peak as an edge
Appropriate preprocessing is necessary for
suppressing noises
Region
growing
Iterative. If neighboring
regions have similar
properties, combine them
Simple Inaccurate due to its local optimization
policy
Clustering Grouping pixels with
similar properties
Simple. Popular clustering
algorithms, such as k-means,
can be used
Difficulty in balancing locational proximity
and pixel value similarity
Active
contour
model
Optimally locating a
deformable closed contour
around a single
target object
Robust by its optimization
framework. If the contour of a
target object is invisible, it still
provides closed contour
Only for a single object. Difficulties of
dealing with unsmooth contours. Usually,
characteristics of the region enclosed by the
contour are not
Considered
Template
matching
and
recognition
based
method
Finding pixels or blocks
whose appearance or other
characteristics are similar to
reference patterns of the
target object
Capable of stable segmentation
by using various pattern
recognition theories for
evaluating the similarity
Computationally expensive. Often a
sufficient number of reference patterns are
necessary for realizing stability
Markov
random
field (MRF)
An integrated method to
optimize the segmentation
result
considering the similarity of
neighboring pixels
Accurate and robust. Highly
flexible and capable of using
various criteria
Computationally expensive. Difficult to
implement
![Page 8: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/8.jpg)
8
which are clinically proven. However, more needs to
be done with ever-growing medical data, called big
data now days, in order to discover hidden
knowledge from the data [14, 31].
Healthcare industry generates huge amount of
data. Intelligent processing of such data can reveal
hidden relationships among the data items which will
help in clinical diagnosis. The growth in usage of
medical image processing can improve quality of
services to reduce death toll and improve health
standards of citizens of a country [9, 28].
Lehmann et al. explored B-Spline interpolation
techniques for medical imaging in order to improve
the quality of images. This has an important utility as
healthcare users need to have good visual perception
of images. Banos Jr, Sehn and Krechel proposed a
service model known as ―Integrated Image Access
and Distributed Processing Service‖ which is a
distributed environment which facilitates radiological
medical personnel to gain access to image processing
features. Matthew J et al. proposed an application for
medical image processing as well as visualization
which enabled professional to study, diagnose
clinical disorders. Later 3D imaging came into
existence to leverage medical image processing. Li,
Papachristou and Shekhar [ provided a reconfigurable
architecture for 3D medical image processing. The
system has four operational stages namely parameter
generation, input brick fetching, medical data stream
processing and output brick storing. Tian and Ha
reviewed applications for medical image processing
that make use of wavelet and inverse transforms.
These applications are used for clinical diagnosis.
Chen, Yi and Ni proposed a platform known as
Medical Image Processing Platform (MIPP), which is
used for web based processing of medical images.
The platform was used for design and manufacturing
of stents used for heart patients [7, 28, 31].
IX. Example for using data mining
techniques to Diagnosis of Cancer & Heart
Ailments
Data mining techniques are being used for
processing medical databases. Kharya proposed a
methodology for diagnosis and prognosis of breast
cancer. Decision tree model was used to represent
actionable knowledge pertaining to breast cancer.
Artificial Neural Networks (ANNs) were also used to
diagnose breast cancer. Krishnaiah et al. used
classification techniques for lung cancer prediction.
Their system is used for early detection of lung
cancer and accurate diagnosis of it which will save
time of doctors besides helping them in clinical
practices. Srinivas et al. used data mining techniques
for prediction of heart attacks [12].
X. Archiving and its Challenges
An archive is a location containing a collection of
records, documents, or there materials of historical
importance. In the context of computers, it is
generally a long-term storage, often on disks and
tapes. Archiving is typically done in a compressed
format so that data are saved efficiently, using less
memory resources and allowing the whole process of
archiving to be executed rapidly. PACS can archive
images for several years: 3-5 years is very common.
PACS storage has inbuilt mechanisms to take care of
disk failures through RAID (Redundant Array of
Independent Disks, also called inexpensive disks).
Depending on the patient load, types of modalities,
and the duration for which the images are to be
stored, the storage size varies from terabytes to
petabytes or even exabytes and zetabytes [23, 24].
A few challenges present themselves in archiving.
A common misconception in archiving is ―my PACS
is DICOM conformant and hence there will be no
interoperability problems.‖ The reality is that every
PACS has its own internal formats to store data and
its inherent proprietary methods to store image
presentation states and key image notes. When a
hospital needs to be migrated a PACS vendor, the
complete earlier data need to be migrated in the
format of the new PACS. This is both time and
money consuming. Part of the problem occurs
because DICOM is in reality a cooperative standard
and not an enforced one and hence has limitations.
Vendors make their own conformance statements,
which may or may not conform to all that is expected
of them and there may be a few gaps and
inconsistencies [16, 22].
Inability of vendors to comply fully with their
conformance statements occurs occasionally. Such
situations arise when a vendor providing a detailed
conformance standard states that interoperability
between their machine and other vendors machine is
the responsibility of the user and not the vendor's.
Similarly, equipment may have conformed to
DICOM standards at testing and installation, but
subsequent non-conformance when the standards
change is not the vendor's responsibility. Data that
are not in conformance with DICOM standards—
where the data are in a format understood only by a
specific vendor—are called ―dirty data‖ [18, 19].
XI. Features of VNA
VNA is an application engine that handles the
data of a vendor and at a fast speed. It is stationed
between the modality and the PACS [Figure 6]. The
imaging data are pushed to VNA directly from the
modality. Thereafter, VNA forwards it to the PACS,
along with the priors. VNA stores the image
presentation states and key image in DICOM format
![Page 9: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/9.jpg)
9
Figure 6: Vendor Neutral Archive (VNA) is stationed
between the modality and PACS. The imaging data is
pushed to VNA directly from the modality.
Thereafter, VNA forwards it to PACS, along with the
priors. VNA stores the image presentation states and
key image in [17].
So what does VNA do? It simply decouples the
PACS and workstations at the archival layer. Let us
take a situation where (a) the modality did not have a
field in the Graphical User Interface (GUI) to permit
data entry for the technologist and (b) the modality
work list was not supported. In this situation, there
could be the possibility that the accession number is
entered into the study description field, and a
compulsory field in the DICOM header in front of the
accession number is left blank. This would cause
problems not only in efficient workflow but also in
retrieval of images in the future [21].
To handle this problem, an application engine was
developed that would check the DICOM header for
any of the non-conformances and automatically
normalize the DICOM tags and additionally, send the
received DICOM file in its original form. Another
issue that one comes across is that of non-
conformance of transfer syntax. Some of the common
ones being JPEG lossless, JPEG 2000 lossless, JPEG
2000, and Implicit VR Little Endian [17].
As the DICOM standard grows, more and more of
syntax are being added, with different vendors using
different ones. It is quite possible that two vendors,
whose systems are expected to be used together, use
different syntax. This problem is handled by
developing an application engine that receives the
data using one kind of syntax and transmits this data
using the syntax of the target system [20].
Finally, the application engine needs to perform
the above two functions at a fast speed. Thus, it is
stationed between the modalities and the PACS and
performs tag morphing and routing. Table 3 outlines
a list of ideal characteristics of VNA and the
advantages of VNA Figure 7.
Figure 7: Vendor Neutral Archive can work as
enterprise archive for all departments like radiology
and cardiology, seamlessly integrating their data.
Table 3 Ideal characteristics and advantages of vendor neutral archive VNA
Ideal characteristics of VNA Advantages of VNA
As per US FDA it is a class one medical device. Increased workflow efficiency (quality), saving time and labor (quickness).
Includes lifecycle image management. Ability to store huge number of images (quantity).
Manages images as well as other related info ,e.g. ,SR,PR,RT objects ,non-DICOM , waveforms ,pdf ,etc .
Allows switching PACS without requiring a complex image/data migration.
Supports open standards. Being able to use the latest hardware technology.
Supports multiple departments, enterprise, and regional architecture.
Effectively controlling data works as an enterprise.
Allows PACS to be interchangeable.
VNA: Vendor Neutral Archive, FDA: Food and Drug Administration, PACS: Picture Archiving and Communications System, SR: Structured Reports, PR: Presentation States, RT: Radiation Therapy, DICOM: Digital Imaging and Communications in Medicine IHE: Integrating the Healthcare Enterprise.
![Page 10: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/10.jpg)
10
XII. Future of Archiving with VNA
In the end, VNA is an archive that has been
developed on an open architecture. It can be easily
migrated, ported to interface with another vendor's
viewing, acquisition, and workflow engine to manage
medical images and related information.
Besides images sourced from radiology, the latest
PACS will allow storage of images from other
sources such as endoscopes, opthalmoscopes,
bronchoscopes, and from the departments of
dermatology, pathology, etc., An emerging term for
such images is ―Visible Lights‖ [17].
XIII. Conclusion
Image processing and informatics image techniques
are helpful to analyze medical image and make
diagnosis of diseases easier. Since a huge number of
techniques have been proposed, and also many
informatics systems which make the appropriate
technique for a specific task is important. For
example, there are many binarization or segmentation
techniques with different properties and therefore we
need to understand what the best binarization or
segmentation technique for the task is and also which
informatics system is better. This paper can be used
for a brief guide for helping the diagnosing of
diseases to be more faster and effective by get the 3Q
factors, quality of processing data, quantity of data
(since MI is a big data source) and the quickness of
processing this data.
As emphasized, medical images are a very difficult
target even for state-of-the-art image processing and
informatics image techniques. Thus, for a specific
task, we may need to develop a new technique. This
will be possible by a collaboration of biologists and
specialists of image processing and informatics
techniques with enough discussion. On the other
hand, a task can be solved easily by an existing
technique or a combination of existing techniques.
We suggest in this paper this new system to find the
diagnosis by implementing two step ( may be more)
in order to get disease diagnosis faster and more
effective, since it is concerned with human life .
Even in this case, it is worth discussing with an
image processing specialist because she/he will help
to choose appropriate techniques. Like biology,
research on image processing and informatics
techniques continues steadily and will make further
progress in accuracy, robustness, versatility,
usability, computational efficiency, etc.
Many biological tasks can use in the future (SIITS)
for fully automatic image analysis. They also can use
future (or even present) informatics image techniques
proving empirically known biological facts and
discovering new biological facts. Again, for
continued progress, mutual collaboration between
biologists and image processing specialists is very
important.
Bibliography:
1. Uchida, Seiichi. "Image processing and
recognition for biological images."
Development, growth & differentiation 55.4
(2013): 523-549.
2. Langs, Georg, et al. "VISCERAL: Towards large
data in medical imaging—Challenges and
directions." Medical Content-Based Retrieval for
Clinical Decision Support. Springer Berlin
Heidelberg, 2013. 92-98.
3. Ravudu, M., V. Jain, and M. M. R. Kunda.
"Review of image processing techniques for
automatic detection of eye diseases." Sensing
Technology (ICST), 2012 Sixth International
Conference on. IEEE, 2012.
4. Richter, Detlef. "Current state of image
processing for medical irradiation therapy."
Radioelektronika (RADIOELEKTRONIKA),
2012 22nd International Conference. IEEE,
2012.
5. Bulsara, Viralkumar, et al. "Low cost medical
image processing system for rural/semi urban
healthcare." Recent Advances in Intelligent
Computational Systems (RAICS), 2011 IEEE.
IEEE, 2011.
6. Villanueva, Lara G., et al. "Medical Diagnosis
Improvement through Image Quality
Enhancement Based on Super-Resolution."
Digital System Design: Architectures, Methods
and Tools (DSD), 2010 13th Euromicro
Conference on. IEEE, 2010.
7. Liu, Danzhou, Kien A. Hua, and Kiminobu
Sugaya. "A generic framework for Internet-based
interactive applications of high-resolution 3-D
medical image data." Information Technology in
Biomedicine, IEEE Transactions on 12.5 (2008):
618-626.
8. Castro, F. Javier Sanchez, et al. "A cross
validation study of deep brain stimulation
targeting: from experts to atlas-based,
segmentation-based and automatic registration
algorithms." Medical Imaging, IEEE
Transactions on 25.11 (2006): 1440-1450.
9. Kassim, Ashraf A., et al. "Motion compensated
lossy-to-lossless compression of 4-D medical
images using integer wavelet transforms."
Information Technology in Biomedicine, IEEE
Transactions on 9.1 (2005): 132-138.
10. Chen, Gong, Hong Yi, and Zhonghua Ni.
"MIPP: a Web-based medical image processing
system for stent design and manufacturing."
Services Systems and Services Management,
![Page 11: 3Q_Factors_to_Enhance_Big_Data_Medical_Image_for_Better_Diagnosis.pdf](https://reader036.fdocuments.in/reader036/viewer/2022081908/563db848550346aa9a924084/html5/thumbnails/11.jpg)
11
2005. Proceedings of ICSSSM'05. 2005
International Conference on. Vol. 2. IEEE, 2005.
11. Abràmoff, Michael D., Paulo J. Magalhães, and
Sunanda J. Ram. "Image processing with
ImageJ." Biophotonics international 11.7 (2004):
36-43.
12. Jannin, Pierre, et al. "Validation of medical
image processing in image-guided therapy."
IEEE Transactions on Medical Imaging 21.12
(2002): 1445-9.
13. de Moraes Barros Jr, Euclides, et al. "A model
for distributed medical image processing using
CORBA." Computer-Based Medical Systems,
2001. CBMS 2001. Proceedings. 14th IEEE
Symposium on. IEEE, 2001.
14. Lehmann, Thomas Martin, Claudia Gonner, and
Klaus Spitzer. "Survey: Interpolation methods in
medical image processing." Medical Imaging,
IEEE Transactions on 18.11 (1999): 1049-1075.
15. Drukker, Karen. "Applied Medical Image
Processing: A Basic Course." Med. Phys 37
.6500 :(2010)
16. van Ooijen, Peter MA, et al. "DICOM data
migration for PACS transition: procedure and
pitfalls." International journal of computer
assisted radiology and surgery (2014): 1-10.
17. Sanjeev, Tapesh Kumar Agarwal. "Vendor
neutral archive in PACS." Indian Journal of
Radiology and Imaging 22.4 (2012): 242-245.
18. Pianykh, Oleg S. Digital imaging and
communications in medicine (DICOM).
Springer, 2012.
19. Suapang, Piyamas, Kobchai Dejhan, and
Surapun Yimmun. "A web-based DICOM-
format image archive, medical image
compression and DICOM viewer system for
teleradiology application." SICE Annual
Conference 2010, Proceedings of. IEEE, 2010.
20. Kaur, K., Chopra, V., & Kaur, H. (2012). Image
Compression of medical images using VQ-
Huffman Coding Technique. International
Journal of Research in Business and Technology,
.36-43 ,(1)1
21. Suapang, Piyamas, Kobchai Dejhan, and
Surapun Yimmun. "Medical image compression
and DICOM-format image archive." ICCAS-
SICE, 2009. IEEE, 2009.
22. Robertson, Ian D., and Travis Saveraid.
"Hospital, radiology, and picture archiving and
communication systems." Veterinary radiology
& ultrasound 49.s1 (2008): S19-S28.
23. Doi, Kunio. "Computer-aided diagnosis in
medical imaging: historical review, current status
and future potential." Computerized medical
imaging and graphics 31.4-5 (2007): 198-211.
24. Liu, Boqiang, et al. "Medical image conversion
with DICOM." Electrical and Computer
Engineering, 2007. CCECE 2007. Canadian
Conference on. IEEE, 2007.
25. Cho, Kyucheol, et al. "Development of medical
imaging viewer: role in DICOM standard."
Enterprise networking and Computing in
Healthcare Industry, 2005. HEALTHCOM 2005.
Proceedings of 7th International Workshop on.
IEEE, 2005.
26. Ryu, Seewon, and Tae-Min Song. "Big Data
Analysis in Healthcare." Healthcare informatics
research 20.4 (2014): 247-248.
27. Raghupathi, Wullianallur, and Viju Raghupathi.
"Big data analytics in healthcare: promise and
potential." Health Information Science and
Systems 2.1 (2014): 3.
28. Bindu, A., and C. N. Kumar. "Inpainting for Big
Data." Signal and Image Processing (ICSIP),
2014 Fifth International Conference on. IEEE,
2014.
29. Koumpouros, Yiannis. "Big Data in Healthcare."
Healthcare Administration: Concepts,
Methodologies, Tools, and Applications:
Concepts, Methodologies, Tools, and
Applications (2014): 23.
30. Groves, Peter, et al. "The ‗big data‘revolution in
healthcare." McKinsey Quarterly (2013).
31. Suresh, K., and M. Rajasekhara Babu. "Towards
on high performance computing of medical
imaging based on graphical processing units."
Advanced Computing Technologies (ICACT),
2013 15th International Conference on. IEEE,
2013.
32. Constantinescu, Liviu, Jinman Kim, and Dagan
Feng. "Integration of interactive biomedical
image data visualisation to Internet-based
Personal Health Records for handheld devices."
e-Health Networking, Applications and Services,
2009. Healthcom 2009. 11th International
Conference on. IEEE, 2009.
33. CHAUHAN, NIKITA K., and SWATI J.
PATEL. "Survey on Medical image processing
areas."
34. Inallou, Mohammad Madadpour, Majid
Pouladian, and Bahman Mehri. "The Application
of Partial Differential Equations in Medical
Image Processing."