Visual interpretation of histopathological images is commonly used for breast cancer detection, but it is a complex task that necessitates extensive experience and skills on the part of pathologists. However, visual interpretation difficulties and a lack of experience may result in missed diagnosis. An increasing need of detecting breast cancer regions in Whole Slide Images (WSI) encourages studies on efficient automatic detection. The second-leading cause of death for women globally is breast cancer. CAD has emerged as a promising solution for reducing errors in breast cancer diagnosis, making it a significant tool in this context. Recent developments in computational pathology, like many other issues in histopathology practice, provide promising and constantly improving tools to overcome these limitations. In this work, we consider the detection of Invasive Ductal Carcinoma (IDC), which is a common sub-type of breast cancer, using three Convolutional Neural Network (CNN) models: AlexNet, DenseNet and VGG. The proposed methods are evaluated on a WSI dataset which consists of 162 whole mount images scanned specimen images from 162 patients diagnosed with IDC. From that, 277,524 patches of size 40× images are extracted. The AlexNet has been used for this problem previously and gave an accuracy of 85%. This research employs a more complex model DenseNet, and a simpler model VGG. The proposed methods yield the best quantitative results for automatic detection of IDC regions in terms of regular accuracy of 90.90% and balanced accuracy of 90.50%. Hence, deep learning approaches are capable for classifying malignancy regions of tissue images for diagnosis support.