Once data on the distribution of time-resolved pressure variations have been
obtained, then a high-resolution tomographic image of optical absorption can be
created. The mathematical analyses and image formation methodologies are beyond
the discussion of this chapter but are detailed in the literature [ 42 – 45 ].
10.6 Hyperspectral Imaging
Hyperspectral imaging (HSI) is a non-invasive multimodal medical imaging
modality that combines reflectance andfluorescence techniques for applications
such as disease diagnosis and image-guided surgery [ 47 – 49 ]. The advantages of
HSI are related to the observation that tissue characteristics such as absorption,
fluorescence, and scattering change during the development and progression of
many diseases. When cells are in different disease states theirfluorescent spectra or
diffuse reflectance spectra change. These spectral variations result because diseased
cells typically have modified structures from normal cells or they undergo different
rates of metabolism. Thus as a disease progresses, further changes in the emission
or reflectance spectra will take place. Because a HSI system captures data for such
tissue characteristics, it can assist in identification and diagnosis of diverse tissue
abnormalities. The applications include identification of various types of cancer,
assessment of diabetic foot ulcers, assessment of tissue wound oxygenation,
mapping of eye diseases, identification of abnormalities in gastrointestinal tissue,
and identification and monitoring of changes in coronary arteries that may be
associated with atherosclerosis.
The HSI system operates by collecting sets of spectral information at continuous
wavelengths at each pixel of a 2D detector array. These data are then analyzed to
determine the absorption spectroscopic characteristics of multiple tissue compo-
nents in a tissue sample. The result is a 3D set of spatially distributed spectral
information that shows where each spectral point is located in the sample. The 3D
dataset is known as ahypercube and is illustrated on the left-hand side in
Fig.10.16. In thisfigure, the curve shows the spectral characteristic of an example
pixel in the 2D slice through the tissue sample.
UV
Continuous scan
pixel by pixel over
a spectral range
from UV to NIRNIR
Spectral information
from one axial scany axisx axisReflectanceFig. 10.16Example of hyperspectral imaging
316 10 Optical Imaging Procedures