Abstract:
Currently in the domain of breast cancer detection, biopsy and ultrasound imaging technique is
being used. Between these two techniques biopsy is costly, time consuming and tissue damaging
for patients. Ultrasound imaging technique is an easy alternative which needs experts to accurately
detect breast lesion and it's type. In the BIRAD method of identi cation, Sonologists look for certain
features of the malignant lesions. These malignant lesions in static image has some signi cant
features which with 15-30% cases result in contradictory results. Most important feature that a
breast lesion posses is it's tissue property ie. tissue destiny and the deforming mechanism. These
properties are associated with the strain value when pressure is applied to the tissue. There are
di erent ways the strain value can be calculated by using multiple images with di erent strain
values. Generally,the malignant lesions exhibits more rigidity than the benign ones. From this
motivation, optical
ow algorithm has been implemented to calculate the strain value and compare
it's measure of mean related to accuracy with cross correlation method used to calculate the strain
values. From this motivation, we have gured out strain with higher accuracy adopting our \M-
Filtered B-Mode Image Tissue Elasticity" estimation method. Higher the accuracy, higher will be
the mean of the calculated strain. Satisfactory SNR (dB) results in our implemented method of
optical
ow which is promising for a classi cation with higher accuracy.
