Introduction
Digital Holographic Interferometric Microscope
Thickness Determination of RBC
Cell Identification
Future Scopes
Conclusion
Malaria is one of the most
widespread and potentially fatal
diseases especially in Africa and Asia
Clinical diagnosis of malaria is based on microscopic
inspection of blood smears by visual inspection of a technician
Much beneficial when automatically discriminable
easy-to-use devices are used instead of visual identification
Use of Interference techniques, Digital Holographic
Microscopy and Interferometric Comparision
Digital holographic microscopy (DHM) is an effective
tool for 3-D imaging of micro-objects
Object phase information is provided by Interferometric
Comparision of phases of the object as well as its
background from the recorded holograms
Digital Holographic Interferometric
Microscope
The location of cells in the field of
view is obtained from the
thickness profile
CELL IDENTIFICATION
Thresholding the thickness distribution by the resolution
of the system, location of cells can be automatically
determined
IDENTIFICATION USING
SINGLE RECONSTRUCTION PLANE
Cell identification using single plane
Top row shows phase-contrast images of four different healthy RBCs.
Bottom row depicts cross-sectional thickness profile along the center line.
Top row shows phase-contrast images of four different malaria-infected RBCs.
Bottom row depicts cross-sectional thickness profile.
Average correlation coefficient from shape comparison of different
cell pairs using data from a single reconstruction plane
(■ healthy, ▲ malaria infected, ------ threshold)
A threshold of 0.88 yielded the best discrimination probability
69% malaria infected cells could be correctly identified when
compared with that of healthy cells (TPR)
FPR is found to be 27%
IDENTIFICATION USING
MULTIPLE RECONSTRUCTION PLANES
Cell identification using multiple planes
Average shape correlation is found at different axial
planes to compute correlation coefficient
Phase-contrast images of a healthy RBC obtained at various axial distances.
Phase-contrast images of a malaria-infected RBC obtained at various axial distances.
Average correlation coefficient from shape comparison of different cell
pairs using data from 20 axial planes (■ healthy, ▲malaria infected, ---threshold)
Probability of correct classification is increased to 84%
with reduced FPR of 11%
Hence use of thickness information at multiple axial planes
will lead to a better probability of identification
ROC curves for the detection of malaria-infected RBCs
Refractive index of blood plasma and RBC could vary from
person to person
Hence a variation of upto 5% is introduced to refractive
index of RBC, plasma and malaria parasites
The probability of discrimination
was found to be 86% and 91% for
constant and correct RI values
respectively
Future of the work lies in using the technique to study other
diseases affecting RBCs
FUTURE SCOPES
Extraction of information along
the focus in a single shot will make
the method faster
A database of healthy and diseased
cells can be made, and a test cell can
be compared with this database to
determine its state of health
By using thickness profile from multiple axial planes,
the recognition performance can be improved
Integration of DHM and correlation algorithms acts as an
automated technique to discriminate different classes of RBCs
Comparison of the shape of the test cell with the database
of healthy and infected cells may indicate whether the cell
is healthy or not.
CONCLUSION
REFERENCES
www.ieeexplore.com
www.google.com
www.howstuffworks.com
www.wikipedia.org
Thank You