Detecting Abnormal Shape
Deformations using Yamabe Flow and Beltrami Coefficents
Project Description:
We address the problem of detecting the region of abnormal
changes on surfaces. It is of great importance in shape analysis, such as
detecting abnormalities on biological shapes. We propose an effective algorithm
to detect the abnormal deformations, by generating quasi-conformal maps
between the original and the deformed surfaces and study its Beltrami coefficient.
We first conformally flatten the surfaces onto the rectangles using Yamabe
flow method and compute the appropriate quasi-conformal map that matches
features. By formulating abnormal changes as non-conformal deformations,
we detect abnormalities by computing the Beltrami coeffient associated uniquely
with the quasi-conformal map. The Beltrami coefficient is a complex-valued
function defined on the surface which describes the conformality of the
deformation at each points. By considering the norm of the Beltrami coefficient,
we can effectively segment the regions of abnormal changes. Furthermore,
by considering the argument of the Beltrami coefficient, we can captures
the rotational changes of the abnormalities. We have tested the algorithm
on synthetic surfaces and real brain surfaces. Experimental results show
that our algorithm can effectively detect the abnormalities and capture their
rotational changes.
Publication:
Result: