A new deep learning classifier. The network consists of the existing DenseNet-121 and the ResNet-50 networks. We observed in our study that in some instances DenseNet and Resnet have orthogonal performances; sometimesDenseNet performed better and sometimes ResNet. In our proposed model, we have created an extralayer with convolutional neural network (CNN) blocks to join these two models together to establish superior performance over either model for all characteristics, namely, recall, precision, AUC-ROC,and F1. The same strategy can be useful in other applications where two competing networks with complementary performance are observed.
Install guidance:
- fork the code
- cd .
- pip instal .
- cite the paper(doi:) https://doi.org/10.1016/j.compmedimag.2021.102008
P.S. python==3.6.3 is recommnded