Image processing from Oaxaca population mammograms to search for breast cancer

What is it about?

Breast density from breast fibroglandular tissue reduces mammography sensitivity. To enhance mammography sensitivity without using x ray, we have proposed using morphology and texture patterns analysis in a specific database of mammograms from Oaxaca population. We propose their analysis using mathematical morphology by means of circularity (κ) and texture using mean height-width ratio of extrema (ρ). These have shown to help as specific descriptors in digital mammograms, and these needs a nonlinear criterion as artificial neural networks to classify benign and malignant histopathology results.

Featured Image

Photo by National Cancer Institute on Unsplash

Photo by National Cancer Institute on Unsplash

Why is it important?

To support mammography, artificial neural network (ANN) or deep learning (DL) have been employed. These techniques using thousands of mammograms have achieved until 99.7% precision for benign/malign mammograms classification. The proposal of this work is to analyze a minor database of specific mammograms from Oaxaca population with a classic approach analyzing morphology and texture patterns. At our knowledge there is no evidence of breast cancer statistical research in Oaxaca population and this work expose some histopathology results of breast cancer incidence. We conclude descriptors proposed have a potential to distinguish benign from malignant condition with the help of a nonlinear criterion of classification.

Perspectives