Name
Abstract | ||
|---|---|---|
In this paper, we propose a compact network called compact unsupervised network (CUNet) to address the image classification challenge. Contrasting the usual learning approach of convolutional neural networks, learning is achieved by the simple K-means on diverse image patches. This approach performs well even with scarcely labeled training images, greatly reducing the computational cost, while maintaining high discriminative power. Furthermore, we propose a new weighted pooling method in which different weighting values of adjacent neurons are considered. This strategy leads to improved classification since the network becomes more robust against small image distortions. In the output layer, CUNet integrates feature maps obtained in the last hidden layer, and straightforwardly computes histograms in nonoverlapped blocks. To reduce feature redundancy, we also implement the max-pooling operation on adjacent blocks to select the most competitive features. Comprehensive experiments on well-established databases are conducted to validate the classification performances of the introduced CUNet approach. |
Year | Venue | DocType |
|---|---|---|
2018 | IEEE Trans. Multimedia | Journal |
Volume | Issue | Citations |
abs/1607.01577 | 8 | 0 |
PageRank | References | Authors |
0.34 | 28 | 6 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Le Dong | 1 | 31 | 7.60 |
| Ling He | 2 | 52 | 6.94 |
| Gaipeng Kong | 3 | 10 | 1.50 |
| Qianni Zhang | 4 | 113 | 24.17 |
| Xiaochun Cao | 5 | 1986 | 131.55 |
| ebroul izquierdo | 6 | 1050 | 148.03 |