Deep Learning for Breast Cancer Diagnosis from Mammograms—A Comparative Study

Tsochatzidis, Lazaros; Costaridou, Lena; Pratikakis, Ioannis

doi:10.3390/jimaging5030037

Cited by 139 publications

(77 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The pre-trained models appear to possess relatively good performance with previous knowledge of large scale data of the existing model. This section describes a background on three popular pre-trained models (VGG-16, GoogleNet and ResNet-50) which have been proven to be excellent in medical image classification [28] [29].…”

Section: Deep Transfer Learning Model Of Cnnmentioning

confidence: 99%

“…The stack of convolutional layers is followed by three fully connected (fc) layers and finally, the softmax layer. The ability of VGG to perform through an increase of effective network receptive field is one of the major advantages due to stacking stage of multiple convolution layers with small kernels as well as limiting the number of parameters [28]. For a number of years, this model in literature has been prevalent to the area of medical image classification such as mammograms [30].…”

Section: A Vgg-16mentioning

confidence: 99%

“…The model is implemented using the strategic approach of stacking nine Inception modules. The implementation presents the advantage of employing optimal local sparse structure using dense components of a multi-layer network [28]. Particularly, the Inception module has included four branches.…”

Section: B Googlenetmentioning

confidence: 99%

See 2 more Smart Citations

Deep Transfer Learning Application for Automated Ischemic Classification in Posterior Fossa CT Images

Suberi¹,

Nurshazwani²,

Tomari³

et al. 2019

IJACSA

View full text Add to dashboard Cite

Computed Tomography (CT) imaging is one of the conventional tools used to diagnose ischemic in Posterior Fossa (PF). Radiologist commonly diagnoses ischemic in PF through CT imaging manually. However, such a procedure could be strenuous and time consuming for large scale images, depending on the expertise and ischemic visibility. With the rapid development of computer technology, automatic image classification based on Machine Learning (ML) is widely been developed as a second opinion to the ischemic diagnosis. The practical performance of ML is challenged by the emergence of deep learning applications in healthcare. In this study, we evaluate the performance of deep transfer learning models of Convolutional Neural Network (CNN); VGG-16, GoogleNet and ResNet-50 to classify the normal and abnormal (ischemic) brain CT images of PF. This is the first study that intensively studies the application of deep transfer learning for automated ischemic classification in the posterior part of brain CT images. The experimental results show that ResNet-50 is capable to achieve the highest accuracy performance in comparison to other proposed models. Overall, this automatic classification provides a convenient and time-saving tool for improving medical diagnosis.

show abstract

Section: Deep Transfer Learning Model Of Cnnmentioning

confidence: 99%

Section: A Vgg-16mentioning

confidence: 99%

See 1 more Smart Citation

Deep Transfer Learning Application for Automated Ischemic Classification in Posterior Fossa CT Images

Suberi¹,

Nurshazwani²,

Tomari³

et al. 2019

IJACSA

View full text Add to dashboard Cite

show abstract

“…In contrast to classical machine learning, which can be applied only to transformed image data, the benefit of deep learning is its ability to process raw image data (LeCun et al, 2015;Moen et al, 2019). This led to increasing importance of deep learning in biology and medicine supporting, first, bioinformatics analysis of protein function and prediction of pathway-related gene function (Le et al, 2018(Le et al, , 2019Al-Ajlan and El Allali, 2019) and, second, diverse image-centred applications, such as segmentation, feature enhancement and recognition, and classification tasks, for optimised workflow in medical diagnosis (Esteva et al, 2017;Kermany et al, 2018;Alom et al, 2019;Tsochatzidis et al, 2019;Black et al, 2020), as well as reconstruction of superresolutional fluorescence images (Weigert et al, 2018;Belthangady and Royer, 2019) or cytometric high content analysis and phenotyping (Scheeder et al, 2018;Yao et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

Application and Comparison of Supervised Learning Strategies to Classify Polarity of Epithelial Cell Spheroids in 3D Culture

et al. 2020

View full text Add to dashboard Cite

Three-dimensional culture systems that allow generation of monolayered epithelial cell spheroids are widely used to study epithelial function in vitro. Epithelial spheroid formation is applied to address cellular consequences of (mono)-genetic disorders, that is, ciliopathies, in toxicity testing, or to develop treatment options aimed to restore proper epithelial cell characteristics and function. With the potential of a highthroughput method, the main obstacle to efficient application of the spheroid formation assay so far is the laborious, time-consuming, and bias-prone analysis of spheroid images by individuals. Hundredths of multidimensional fluorescence images are blinded, rated by three persons, and subsequently, differences in ratings are compared and discussed. Here, we apply supervised learning and compare strategies based on machine learning versus deep learning. While deep learning approaches can directly process raw image data, machine learning requires transformed data of features extracted from fluorescence images. We verify the accuracy of both strategies on a validation data set, analyse an experimental data set, and observe that different strategies can be very accurate. Deep learning, however, is less sensitive to overfitting and experimental batch-to-batch variations, thus providing a rather powerful and easily adjustable classification tool.

show abstract

“…So, especially after the recent advances in deep neural networks (DNN), improvements in biomedical images analysis could be exploited to enhance the performance of CAD [7][8]. In following we will cite some recent studies:…”

Section: Introductionmentioning

confidence: 99%

A New Triplet Convolutional Neural Network for Classification of Lesions on Mammograms

Medjeded¹,

Mahmoudi²,

Chenine³

et al. 2019

RIA

View full text Add to dashboard Cite

Mammography provides a useful tool for breast cancer detection. However, many doctors have difficulty in making the right decision based on mammograms. This paper aims to set up a deep learning (DL) architecture that can effectively differentiate between benign and malignant tumors. Specifically, a new triplet convolutional neural network (CNN) was established with three subnetworks, each of which contains a succession of layer blocks. Each block consists of two convolutional layers, a dropout layer and a max-pooling layer. During operation, the region of interest (ROI) extracted from the mammogram is imputed to the first subnetwork, and processed by the Canny filter. The filtered results become the input of the second subnetwork, while the third subnetwork takes the whole image as input. To verify the effectiveness of our architecture, a set of 500 images from 301 patients was extracted from the DDSM database and augmented to 4,000 images, and divided into a training set (80 %) and a testing set (20 %). The results show that our architecture achieved an accuracy of 93.13 %, a sensitivity of 96 % and a specificity of 90.25 %. This research provides a desirable way to identify breast cancer based on mammography.

show abstract

Deep Learning for Breast Cancer Diagnosis from Mammograms—A Comparative Study

Cited by 139 publications

References 21 publications

Deep Transfer Learning Application for Automated Ischemic Classification in Posterior Fossa CT Images

Deep Transfer Learning Application for Automated Ischemic Classification in Posterior Fossa CT Images

Application and Comparison of Supervised Learning Strategies to Classify Polarity of Epithelial Cell Spheroids in 3D Culture

A New Triplet Convolutional Neural Network for Classification of Lesions on Mammograms

Contact Info

Product

Resources

About