Methods for Image Registration and Image Registration Validation in
Colposcopy Images

		       Juan David Garcia
		       garcia@fel.cvut.cz

Uterine Cervical Cancer is the second most common cancer among women
worldwide. It is a slow growing cancer with high cure rates when
detected and treated early.

Presently, the combination of colposcopy, a diagnostic method used to
visually detect cancerous and pre-cancerous tissue in the cervix, with
Computer-Aided Diagnosis (CAD) and multispectral imaging is one of the
most promising lines of research for improved screening and early
detection. In this system, when multiple image analysis is involved,
image registration must be used to ensure pixel correspondence to the
same tissue location on all analyzed images. We present the results of
our research on this subject: We have developed an elastic algorithm
for cervical image registration and, in parallel, we have developed a
feature based system to generate landmarks which may be used to test
the registration validation.