Abstract
In this paper, a new optimization approach is designed for convolutional neural network (CNN) which introduces explicit logical relations between filters in the convolutional layer. In a conventional CNN, the filters’ weights in convolutionallayers are separately trained by their own residual errors, and the relations of these filters are not explored for learning. Different from the traditional learning mechanism, the proposed correlative filters (CFs) are initiated and trained jointly in accordance with predefined correlations, which are efficient to work cooperatively and finally make a more generalized optical system. The improvement in CNN performance with the proposed CF is verified on five benchmark image classification datasets, including CIFAR-10, CIFAR-100, MNIST, STL-10, and street view house number. The comparative experimental results demonstrate that the proposed approach outperforms a number of state-of-the-art CNN approaches.
| Original language | English |
|---|---|
| Pages (from-to) | 3218-3229 |
| Journal | IEEE Transactions on Cybernetics |
| Volume | 47 |
| Issue number | 10 |
| Early online date | 13 Dec 2016 |
| DOIs | |
| Publication status | Published - Oct 2017 |
| Externally published | Yes |
Bibliographical note
This work was supported in part by the National Natural Science Foundation of China under Grant 61472281 and Grant 61622115, and in part by the Program for Professor of Special Appointment (Eastern Scholar) at the Shanghai Institutions of Higher Learning under Grant GZ2015005.Keywords
- Convolutional kernel
- convolutional neural network (CNN)
- correlative filters (CFS)
- filter modeling
- image classification