Neural network based classification of mammographic microcalcification for breast cancer diagnosis / Nina Korlina Madzhi, Mahanijah Md Kamal and Rosni Abu Kassim
Initially, mammography images of the breast were produced using standard x-ray machine, but today, the breast is imaged on state-or-the-art machines, capable of producing fine detail with minimal exposure to the patient. In this paper, the performance of recently developed neural network structure,...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Research Reports |
| Language: | English |
| Published: |
2006
|
| Subjects: | |
| Online Access: | http://ir.uitm.edu.my/id/eprint/47688/1/47688.pdf http://ir.uitm.edu.my/id/eprint/47688/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Initially, mammography images of the breast were produced using standard x-ray machine, but today, the breast is imaged on state-or-the-art machines, capable of producing fine detail with minimal exposure to the patient. In this paper, the performance of recently developed neural network structure, General Regression Neural Network (GRNN), was examined on the on-line Database for Screening Mammography of University of South Florida (DDSM). This is a well used database in machine learning, neural network and image processing. They are commonly used to increase the accuracy . of breast cancer diagnosis. Tn this study, first we have to carry out a preprocessing step which consists to remove or attenuate the curvilinear structures present in a mammogram and corresponding to the blood vessels, veins, milk ducts, speculations and fibrous tissue. Then the gradient of the preprocessed image is calculated and finally the data from three classes of digital mammography images were used for training and testing the approximation function of GRNN structure. The three classes consist of normal, benign and cancer cases. The accuracy performances of the three classes were achieved by using the spread value of 1.2 for each class. |
|---|