Abstract
Convolutional neural networks have significantly improved the field of computer vision. Most of the standard deep learning models for vision are based on convolutional neural networks. One such model is DenseNet, which is well known for providing comparable results to ResNet with fewer trainable variables. This paper proposes a mechanism to improve the performance of DenseNet architecture. We have compared our proposed approach with the original DenseNet architecture on CIFAR100 dataset. Effects of applying dropout mechanism with our proposed mechanism is also studied on DenseNet model and the obtained results showed that our approach helps DenseNet to significantly improve the performance and helping the model to learn faster. Even in situations where original DenseNet-121 would have overfit, our mechanism helped DenseNet-121 to keep improving the accuracy without overfitting on the dataset assuring better accuracy on testing data.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.