Automated Detection of Benign and Malignant in Breast Histopathology Images

Breast cancer detection and classification using histological images play a critical role in the breast cancer diagnosis process. This paper presents a framework for autodetection and classification of breast cancer from microscopic histological images. The images are classified into benign or malig...

Full description

Saved in:
Bibliographic Details
Main Authors: Baker, Qanita Bani, Zaitoun, Toqa' Abu, Banat, Sajda, Eaydat, Eman, Alsmirat, Mohammad
Format: Conference Proceeding
Language:English
Subjects:
Benign
Breast cancer
Clustering algorithms
Digital Pathology
Feature extraction
Image Analysis
Image color analysis
Image segmentation
K-means
Malignant
Microscopic Images
Watershed
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Breast cancer detection and classification using histological images play a critical role in the breast cancer diagnosis process. This paper presents a framework for autodetection and classification of breast cancer from microscopic histological images. The images are classified into benign or malignant. The proposed framework involves several steps which include image enhancement, image segmentation, features extraction, and images classification. The proposed framework utilizes a novel combination of K-means clustering and watershed algorithms in the segmentation step. We used K-means clustering to produce an initial segmented image and then we applied the watershed segmentation algorithm. Classification results show that the proposed method effectively detect and classify breast cancer from histological image with accuracy of 70.7% using a proposed Rule-Based classifier and 86.5% using a Decision Tree classifier.
ISSN:2161-5330
DOI:10.1109/AICCSA.2018.8612799