11 Remote Space- and Time-Resolved Skin Perfusion Detection …
193
is not constant and/or homogeneous but instead, consists of a complex interaction
of neighboring skin areas [15]. Further examples for functional mapping of PPGI
sequences are presented in Chap. 1.
11.4.3 Additional Analysis of Morphologic Features
In addition to an analysis of the temporal devolvement of pixel or ROI intensities, the
spatial picture content can be analyzed directly. For this, morphologic features can
be extracted from single or a sequence of PPGI frames, thereby providing additional
information about the recorded sequence or the monitored subject, which can then
be used for the following:
• segmentation of image content (e.g., distinguishing between background and
foreground or recognition of body parts),
• recognition of body posture,
• detection of gestures, or
• detection of facial expressions.
Depending on the distinct application, several methods of digital image processing
are available [17–20]. As such, the individual algorithms are not considered in detail
here.
11.5 Detection and Compensation of Movement Artifacts
Due to the remote sensing concept of the PPGI, in general, movements of the
measured object cannot be avoided totally. In contrast to (adhesive) skin-contact
sensors which receive bodily signals always from the tissue directly below them,
the pixel sensors in the camera setup may be assigned to different skin regions
if the measured object is moving relative to the camera. In this way, changes of
detected brightness are not only caused by the tissue properties under investigation
but instead, can also arise from varying brightness levels of different skin surfaces
which are recorded from the same pixel over time. Depending on the measurement
site and the type and amplitude of the motion, it is possible that such movement
artifacts can totally mask the vital signals in the recorded data.
Consequently, the optimum approach for PPGI recordings would be to prevent
such movement right from the beginning, particularly for recordings under laboratory
conditions, in which case the object under study should be fixed or at least positioned
appropriately such that potential movements are prevented or minimized. Since the
use of fixations places considerable constraints on possible applications, movement
maneuvers should be detected and overcome by means of digital image and/or video
processing otherwise.
193
is not constant and/or homogeneous but instead, consists of a complex interaction
of neighboring skin areas [15]. Further examples for functional mapping of PPGI
sequences are presented in Chap. 1.
11.4.3 Additional Analysis of Morphologic Features
In addition to an analysis of the temporal devolvement of pixel or ROI intensities, the
spatial picture content can be analyzed directly. For this, morphologic features can
be extracted from single or a sequence of PPGI frames, thereby providing additional
information about the recorded sequence or the monitored subject, which can then
be used for the following:
• segmentation of image content (e.g., distinguishing between background and
foreground or recognition of body parts),
• recognition of body posture,
• detection of gestures, or
• detection of facial expressions.
Depending on the distinct application, several methods of digital image processing
are available [17–20]. As such, the individual algorithms are not considered in detail
here.
11.5 Detection and Compensation of Movement Artifacts
Due to the remote sensing concept of the PPGI, in general, movements of the
measured object cannot be avoided totally. In contrast to (adhesive) skin-contact
sensors which receive bodily signals always from the tissue directly below them,
the pixel sensors in the camera setup may be assigned to different skin regions
if the measured object is moving relative to the camera. In this way, changes of
detected brightness are not only caused by the tissue properties under investigation
but instead, can also arise from varying brightness levels of different skin surfaces
which are recorded from the same pixel over time. Depending on the measurement
site and the type and amplitude of the motion, it is possible that such movement
artifacts can totally mask the vital signals in the recorded data.
Consequently, the optimum approach for PPGI recordings would be to prevent
such movement right from the beginning, particularly for recordings under laboratory
conditions, in which case the object under study should be fixed or at least positioned
appropriately such that potential movements are prevented or minimized. Since the
use of fixations places considerable constraints on possible applications, movement
maneuvers should be detected and overcome by means of digital image and/or video
processing otherwise.