Image Denoising Using Wavelets: Application in Medical Imaging - Advances in Heuristic Signal Processing and Applications - page 300

Fig. 13.16 Representation of the interrelated components of the imaging sequence

and optics (coherent optical tomography, diffuse optical imaging). These various
image sources provide usually complementary insights, whether it be in the mor-
phological (structure and anatomy) or functional (physiology, metabolism) plane.
Figure 13.16 gives an overview of medical image formation and exploitation:

• An object corresponds to an organ or a lesion (bone fracture, tumor, atheroma-

tous plaque). It may be static or moving, a superficial tissue element or a large,
important structure;

• The Sensor section corresponds to the technique of observation or measure con-

cerned which allows the exploration of this object in its context, or separately, at
a macroscopic level just as at the microscopic level, its shape, its architecture, or
while functioning;

• Data (2D, 3D, or time sequence images) are only obtained after using signal

processing techniques and reconstruction algorithms allowing the correction of
defects intrinsic to the sensors (noise, distortion) and to multiple wave-matter in-
teractions. In terms of information management of these data, the DICOM norm
is generally adopted. Methods of signal and image processing, such as noise re-
duction (or denoising) by multiresolution analysis, are indispensable for efficient
processing. However, artifacts frequently remain, and are linked to the reconstruc-
tion methods themselves, as well as to movements of the objects in the course of
acquisition. The final spatiotemporal resolution as well as the contrast between
objects depends on all these factors. Access to an image is often accompanied by
the injection of agent or contrast products (enhancement of vascular structures
in X-ray imaging or even in echography), biomarkers for highlighting lesions or
specific biological processes, and radio-pharmaceutical compounds (tracers) that