Shuffled Shepherd Deer Hunting Optimization based Deep Neural Network for Breast Cancer Classification using Breast Histopathology Images. (May 2023)

• Record Type: Journal Article
• Authors: Bhausaheb, Deshmukh Pramod; Kashyap, Kanchan Lata
• Is Part Of: Biomedical signal processing and control. Volume 83(2023)
• Publication Date: 2023-05
• Subjects: Deep Neural network -- Breast cancer classification -- Histopathology image analysis -- Shuffled Shepherd Deer Hunting Optimization -- Deer Hunting Optimization Algorithm; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2023.104570
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Accurate classification of breast cancer from the histopathology images poses a difficult task because of various benign breast tissue proliferative lesions and heterogeneity of abnormal cell growth. Various breast cancer classification methods are adopted in recent decades to categorize the breast cancer from histopathology images, but generating accurate classification result poses a complex task in medical image analysis system. Therefore, an accurate breast cancer classification method named Shuffled Shepherd Deer Hunting Optimization-based Deep Neural Network (SSDHO-based DNN) is designed for classifying the breast tumor images into six different classes, like non-tubule, non-tumor nuclei, tubule, apoptosis, tumor nuclei, and mitosis. The proposed algorithm named SSDHO is modelled by merging the Shuffled Shepherd Optimization Algorithm (SSOA) and Deer Hunting Optimization Algorithm (DHOA). Here, the feature, like statistical features, shape features and Convolutional Neural Network (CNN) features are effectively mined from the segmented blood cells such that these extracted features make the classification process more effective using DNN. By employing DNN, the breast cancer classification process is achieved more efficiently with their associated hidden neurons and generates the classification result more accurately. The devised approach obtained maximum results in terms of accuracy, precision, sensitivity, and specificity by acquiring the values of 0.9561, 0.8232, 0.7903, and 0.9426, respectively.