DRNet: Dissect and Reconstruct the Convolutional Neural Network via Interpretable Manners

no code implementations • 20 Nov 2019 • Xiaolong Hu, Zhulin An, Chuanguang Yang, Hui Zhu, Kaiqaing Xu, Yongjun Xu

For VGG16 pre-trained on ImageNet, our method averagely gains 14. 29\% accuracy promotion for two-classes sub-tasks.