A New Design for Accurate Breast Cancer Detection and Classification Diagnosis Model based on Feature Extraction with Artificial Neural Networks

Authors

  • Mr. K. Vignesh Kumar, Dr. N. Sumathi

Abstract

In present scenario, Breast Cancer most common and fatal disease that causes high death rates on women worldwide. Constant research works are being done for developing several efficient methods for earlier and accurate cancer detection. However, some typical disease features, such as, micro-calcification and mass that are found on mammograms, are used to enhance the diagnosis results. Traditional techniques need the help of oncologists for examining the breast lesions for cancer detection, which is inefficient and time consuming. Hence, there is a requirement for automated cancer detection model that produces higher rate of accuracy in diseases classification. With that concern, this paper develops a model called Accurate Breast Cancer Detection and Diagnosis (ABCDD). In this, the input patient mammograms are given for pre-processing, in which, noise and artefact removal is done. Further, Region of Interest (ROI) of processed images is segmented with Sobel operator-based edge detection with Adaptive Weiner Filter. And, Advanced Feature Extraction (AFE) operations are carried out by considering the three significant image features, namely, Shape, Intensity and Texture of segmented mammograms. Finally, the images are classified using Artificial Neural Networks (ANN), which are trained and tested for classifying Benign and Malignant images. The obtained results show that the proposed model produces accurate results in cancer image classification in minimal time than other compared works.

Published

2021-09-22

How to Cite

Mr. K. Vignesh Kumar, Dr. N. Sumathi. (2021). A New Design for Accurate Breast Cancer Detection and Classification Diagnosis Model based on Feature Extraction with Artificial Neural Networks. Drugs and Cell Therapies in Hematology, 10(3), 79–96. Retrieved from https://www.dcth.org/index.php/journal/article/view/406

Issue

Section

Articles