PROJECT TITLE :

Her2Net A Deep Framework for Semantic Segmentation and Classification of Cell Membranes and Nuclei in Breast Cancer Evaluation - 2018

ABSTRACT:

We tend to gift an economical deep learning framework for identifying, segmenting, and classifying cell membranes and nuclei from human epidermal growth factor receptor-a pair of (HER2)-stained breast cancer images with minimal user intervention. This can be a long-standing issue for pathologists because the manual quantification of HER2 is error-prone, pricey, and time-consuming. Hence, we propose a deep learning-based mostly HER2 deep neural network (Her2Net) to solve this issue. The convolutional and deconvolutional components of the proposed Her2Net framework consisted mainly of multiple convolution layers, max-pooling layers, spatial pyramid pooling layers, deconvolution layers, up-sampling layers, and trapezoidal long short-term memory (TLSTM). A absolutely connected layer and a softmax layer were also used for classification and error estimation. Finally, HER2 scores were calculated based mostly on the classification results. The most contribution of our proposed Her2Net framework includes the implementation of TLSTM and a deep learning framework for cell membrane and nucleus detection, segmentation, and classification and HER2 scoring. Our proposed Her2Net achieved 96.sixty fourpercent precision, ninety six.seventy ninep.c recall, 96.seventy one% F-score, ninety three.08% negative predictive price, ninety eight.thirty three% accuracy, and a six.84p.c false-positive rate. Our results demonstrate the high accuracy and wide applicability of the proposed Her2Net in the context of HER2 scoring for breast cancer evaluation.