Abstract:
Modulus imaging has great potential in soft-tissue characterization since it reveals intrinsic
mechanical properties. The elastic properties of biological tissues are usually
modified by disease. The physical quantities that describe tissue elastic properties are
stress, strain and elastic moduli. Stress distribution reconstruction is vital to find out
the true value of the modulus distribution. Stress distribution not only depends on the
physiology of the tissue but also largely changes with the boundary conditions. In
clinical practice it is impossible to find out in-depth information about the stress distribution
of the tissue. Stress distribution only can predict based on different parameters.
In this dissertation stress distribution is predicted from the strain distribution and stress
value on the top surface of the tissue body. For predicting the first stress distribution
the difference of strain distributions and the top surface stresses are taken as the references.
The stress distribution prediction is updated by addressing the error of the
predicting strain values. Simulation is carried out for different aspects of the tissue
background and the inclusion. The proposed concept is validated with the recent algorithm
by the simulation data first then comparing with the recently published data.
The modulus prediction error is around 4% to 5.5% comparing with the real modulus
value. Researchers in this field show the inclusion to background modulus mean error
as 6.87% for a particular phantom whereas the proposed algorithm found the mean
error 4.69% for the same kind of phantoms. The overall stress prediction will improve
the modulus prediction which refers to good quality modulus image. Clinically it will
provide detail information about the modulus spectrum.
