On-Chip Holographic Microscopy and its Application for Automated Semen Analysis - Biomedical Optical Phase Microscopy and Nanoscopy - page 159

(A) (B)

100 µ m 100 µ m

Figure 8.3
(A) A digitally subtracted lensless hologram shows the motion of eight sperms moving between
two successive frames (500 ms apart). (B) A microscopic image is digitally reconstructed from the
lensless differential hologram shown in (A) to illustrate the positions of eight sperms in two
successive frames (black spots indicate the sperms' starting positions and white spots indicate
their end positions). The displacement vectors of these sperms are labeled with green arrows. For
interpretation of the references to color in this figure legend, the reader is referred to the web
version of this topic.

must have an adequate length and an orientation aligned with its head to qualify for a
positive count. To ensure this, a determinant-of-Hessian filter [73] is applied to enhance the
ridge-like features of sperm tails before matching the enhanced ridges to the location of the
detected heads. If the length of a tail ridge is outside a typical range (20 60 μ m) or its
orientation is not aligned with the head, this candidate object will be discarded as negative.

Quantification of motile sperm in semen is a relatively easier task for our technique because
motile sperm are the only moving objects within semen samples. Consecutive lensless
holographic frames are digitally subtracted from each other to create differential holograms
representing the displacement of the sperm ( Figure 8.3A ). These lensfree differential
images are then reconstructed with the same reconstruction algorithm as discussed earlier.
As shown in Figure 8.3B , the displacement of each sperm generates one negative (dark) and
one positive (bright) spot in these reconstructed holographic images, which respectively
represents the start and the end positions of the sperm's motion. As a result, we can
simultaneously estimate the speed and the dynamic trajectories of all the motile sperm
within our large FOV ( B 24 mm 2 ) by quantifying the relative distances between these bright
and dark spots. To do this, the locations of these spots were first detected by simple
thresholding and then refined by their centers of gravity [74] . After finding the nearest
neighbors with opposite polarity for the bright/dark spots, the displacement of individual
sperm between frames is recorded for each bright/dark spot pair (see Figure 8.3B ). One can
next link up these displacements of motile sperm in all consecutive frames and then plot out
the dynamic trajectories of all the sperm within the imaging FOV ( Figure 8.4A ). The
average speed of each motile sperm and the speed distribution of the whole population can