Biomedical Engineering Reference
In-Depth Information
1. imageprocessing/clinical data storage for detection, visualization and
qualitative/quantitative evaluation of abnormalities (
Investigation features
,pic-
tures with red contours in Fig.
12.1
);
2. comparisons among images and clinical data of the same patient (
Follow-up
features
, pictures with orange contours in Fig.
12.1
);
3. quantitative evaluation and retrieval of clinical cases similar to those of unknown
lesions (
Tracking feature
, pictures with blue contours in Fig.
12.1
).
12.3.1 Investigation Features
These features are grouped in two main subjects: recognition and diagnostic. The
first group is related to the exploitation of medical image data, the recognition of
anatomical elements and their visualization. The second group offers all the features
useful for the diagnosis process.
12.3.1.1 Recognition Features
This section describes all those features extracted in medical image analysis with a
CAD system that provides a qualitative view of all anatomical structures (healthy
and diseased) and their location and shape in the patient's body.
Medical image data
Medical image data are usually represented as a stack of individual images. Each
image represents a thin slice of the scanned body part and is composed of individual
pixels
(picture elements). These pixels are arranged on a 2D grid, where the distance
between two pixels is typically constant in each direction.
Volumetric data combine individual images into a 3D representation on a 3D grid.
The data elements are now called
voxels
(volume elements), and they are located on
the grid points.
In medical imaging, any of the various types of equipment or probes used to
acquire images of the body, such as radiography, CT, US or MRI are called
modal-
ities
. Each modality is based on different physical phenomena and thus captures
different types of information. Independently from the image acquisition technique
employed, medical image data are physically stored together with the information
that is essential for the interpretation of the images, like patient's and investigator's
data, image number, image position, image resolution, acquisition time and modality
and scan parameters.
This information is highly standardized as a result of dedicated and long-termstan-
dardization activities. These have led to the Digital Imaging and Communications in
Medicine standard (DICOM), which was established by the National Electrical Man-
ufacturers Association (NEMA). The current version, 3.0, was established in 1993.
DICOM is the industry standard for the transfer of radiological images and other
medical information between computers and medical devices. DICOM enables digi-
