exposure should not exceed 100 μs. If the image of the cells is
saturated (overexposed), either the illumination levels or the
exposure time can be reduced. Conversely, if the image is
underexposed, the best remedy is to increase illumination
levels, if possible; the exposure time should not be increased
beyond the upper bound specified here. If illumination levels
and exposure time have both been maximized, but the image is
still too dark, then one may use any available sensitivityenhancing functionality specific to the camera model (e.g.,
preamplifier gain increase, or output of the analog-to-digital
converter’s lower bits) or, as a last resort, apply digital image
processing tools such as brightness increase or gamma
correction.
56. To unambiguously identify intracellular ice formation, at least
two images of the ice crystal interface must be captured as it
advances through the cell interior. Therefore, the velocity of
intracellular crystal growth, when divided by one-half the smallest diameter of the projected cell image, yields a lower bound
for the image acquisition rate that is required to detect intracellular ice formation. Thus, for a crystallization velocity on the
order 10 μm/ms, and a projected cell diameter of ~10 μm, the
minimum frame rate is ~2000 fps. In practice, there is also an
upper bound on the image acquisition rate. For example, the
frame rate cannot exceed the inverse of the camera exposure
time (set in step 4 of Subheading 3.4), because the exposure of
one frame must be completed before the next frame can be
exposed. Similarly, high-speed cameras typically have a limiting
data throughput rate, so that the maximum possible frame rate
will be determined by the ROI dimensions that were specified
in step 3 of Subheading 3.4. Furthermore, increasing the
frame rate will decrease the available video recording duration,
which may in some cases result in loss of data. On the other
hand, to detect intracellular ice formation, there is typically no
advantage to having more than 5–10 sequential video frames
depicting the growth of the intracellular ice. Therefore, for
cells that have a projected diameter on the order ~10 μm, the
image acquisition rate does not need to be faster than
~8000 fps.
57. Modern high-speed imaging systems save the acquired images
in a circular (first-in-first-out) memory buffer and terminate
the recording after a preset number of frames have been
acquired following detection of the trigger signal. Thus,
some portion of the video that is retained in the camera memory will have been acquired prior to the trigger event, whereas
the post-trigger frames all represent images that were acquired
after the trigger event. In the data analysis procedure described
High-Speed Video Cryomicroscopy
257
saturated (overexposed), either the illumination levels or the
exposure time can be reduced. Conversely, if the image is
underexposed, the best remedy is to increase illumination
levels, if possible; the exposure time should not be increased
beyond the upper bound specified here. If illumination levels
and exposure time have both been maximized, but the image is
still too dark, then one may use any available sensitivityenhancing functionality specific to the camera model (e.g.,
preamplifier gain increase, or output of the analog-to-digital
converter’s lower bits) or, as a last resort, apply digital image
processing tools such as brightness increase or gamma
correction.
56. To unambiguously identify intracellular ice formation, at least
two images of the ice crystal interface must be captured as it
advances through the cell interior. Therefore, the velocity of
intracellular crystal growth, when divided by one-half the smallest diameter of the projected cell image, yields a lower bound
for the image acquisition rate that is required to detect intracellular ice formation. Thus, for a crystallization velocity on the
order 10 μm/ms, and a projected cell diameter of ~10 μm, the
minimum frame rate is ~2000 fps. In practice, there is also an
upper bound on the image acquisition rate. For example, the
frame rate cannot exceed the inverse of the camera exposure
time (set in step 4 of Subheading 3.4), because the exposure of
one frame must be completed before the next frame can be
exposed. Similarly, high-speed cameras typically have a limiting
data throughput rate, so that the maximum possible frame rate
will be determined by the ROI dimensions that were specified
in step 3 of Subheading 3.4. Furthermore, increasing the
frame rate will decrease the available video recording duration,
which may in some cases result in loss of data. On the other
hand, to detect intracellular ice formation, there is typically no
advantage to having more than 5–10 sequential video frames
depicting the growth of the intracellular ice. Therefore, for
cells that have a projected diameter on the order ~10 μm, the
image acquisition rate does not need to be faster than
~8000 fps.
57. Modern high-speed imaging systems save the acquired images
in a circular (first-in-first-out) memory buffer and terminate
the recording after a preset number of frames have been
acquired following detection of the trigger signal. Thus,
some portion of the video that is retained in the camera memory will have been acquired prior to the trigger event, whereas
the post-trigger frames all represent images that were acquired
after the trigger event. In the data analysis procedure described
High-Speed Video Cryomicroscopy
257