Clasificación de lesiones tipo distorsión arquitectural en imágenes de mamografías

Abstract

With the advancement of technologies, Computer
Assisted Diagnostics (CAD) systems have gained a
great boom because they use the power of current
computer tools to integrate different processes in
order to reduce time and cost. This work focuses on
the automatic detection of breast cancer through
digital mammography, focusing on an injury that by
its nature is difficult to detect (architectural
distortion), it exposes the methodology used,
analysis of techniques for the different stages, as
well as the progress that has been made. It shows
results of the interface design, for the manipulation
of the images and the first three stages. During the
preprocessing phase the Seam Carving algorithm is
applied in which you can modify the image size
without altering its quality. At the same stage, a
median filter was selected for edge smoothing. To
reduce noise, the Gabor filter is used which
provides satisfactory results, the insulation of the
lesion.