Improve differentiation of breast mass using fuzzy segmentation method

الملخص EN

Breast cancer is one of the almost public types of cancer in women.

Breast masses generally classified to cystic and solid masses and in both, there are many subtypes.

So that understanding exact types of them is useful for better treatment of the patient.

There are clinic invasive techniques for that but image processing usually seems helpful to get same or better results.

Furthermore, an intelligent computer-aided diagnosis system or proper algorithm can be very helpful for radiologist in detecting and diagnosing abnormal cases earlier and faster than typical screening methods.

One of the most important image processing is Edge detection which targets towards image understanding.

This study tried to improve the early detection of breast masses (cystic/solid) from 2D ultrasound scanning, by running the resulted images of the abnormal cases in image enhancement, edge detection, furthermore comparing between the resulted 2D images by 3D images, Elastography, and then thermal detection of breast mass images.

Four cases of different abnormalities taken from 783 patients had been scanned in Harer hospital by 2D ultrasound B- Mode Linear probe with 7.2 MHz frequency, 47 of them diagnosed as breast mass.

The images were enhanced by adaptive histogram equalization (AHE) which usually given proper contrast enhancement.

Then appropriate conditions were chosen for Fuzzy set theory to give us proper edge detection.

The result was the obtained 2D images are much more clear for diagnosis after running into image processing, the mass easily can be detected and classified into the correct type without using 3D high-cost ultrasound.