Classification of Breast Masses Using Convolutional Neural Network as Feature Extractor and Classifier

Breast cancer detection through attention based feature integration model

“…The deep-net model is denoted by −> β , here α denotes the image and β represents the classification mechanism. A neural network consists of H convolutions and P connected layers shown in (4). Here {g x (. )}…”

“…An X-ray image of decreased dose from the breast of the patient is a mammogram and this is normally captured by the use of two views, namely, medio-lateral oblique (MLO) and bilateral craniocaudal (CC). Over the last two decades, there has been an increase in this methodology that helps radiologists to detect anomalies [4]. Hence, radiologists study the possibility of mass being present, which could be cystic, lumps, or little deposits of calcium that are normally seen in irregular forms and are termed micro-calcification.…”

Multi-Tasking U-Shaped Network for Benign and Malignant Classification of Breast Masses

“…In terms of CNN structure improvement, Jiao et al, removed the fully connected layers used for classification in CNN, and sent the middle-level features and high-level features to two linear support vector machines for BM classification of breast masses [26]. Sarkar et al, used Leaky ReLU as the activation function and applied dropout to the fully connected layer to build a CNN model with four convolutional layers and three fully connected layers for the BM classification [27]. Rampun et al, removed the local response normalization layer on the basis of AlexNet, and added a BN layer after each convolution layer, using the PReLU activation function instead of the ReLU activation function to obtain better classification performance than the original AlexNet Effect [28].…”

“…In CNN-based methods, the pre-processing of model input data is mostly implemented using simple normalization operations [25]- [27]. However, the gray distribution corresponding to breast masses is mainly concentrated in the region with a gray level of more than 10,000, and the gray distribution will not change if a linear normalization operation is performed directly on the mammography.…”

Hybrid computer aided diagnostic system designs for screen film mammograms using DL‐based feature extraction and ML‐based classifiers

“…DL-based CAD system designs have also been popularly used for classification of mammographic masses using SFM images (Abbas, 2016;Abdelhafiz et al, 2019;Aboutallib et al, 2018;Agnes et al, 2019;Al-masni et al, 2018;Khamparia et al, 2021;Kumar et al, 2017;Lecon et al, 2015;Ravishankar et al, 2016). Various traditional ML-based CAD system designs have used hand crafted features, that is, textural and the morphological features (Jehlol et al, 2015;Nock & Nielsen, 2004;Sadad et al, 2018;, while deep learning-based CAD system designs are based on different architectures of convolutional neural networks (Rouhi et al, 2015;Sarkar et al, 2018;Simonyan et al, 2015). The convolutional neural network (CNN)-based deep learning systems are performing very well and becoming omnipresent almost in every large computational image processing system.…”