Name
Title | ||
|---|---|---|
HCKBoost: Hybridized composite kernel boosting with extreme learning machines for hyperspectral image classification. |
Abstract | ||
|---|---|---|
•A new boosting based algorithm which enables to build composite kernels (CKs) by using spatial and spectral hybrid kernels.•We have aimed to develop automated spatial and spectral hybrid kernel construction using different type of kernel transfer functions.•Integrating the contextual and spectral information together is realized with composite kernels (CKs) using convex combination strategy.•The proposed method is a pure ensemble method, therefore, complicated optimization procedures on the multiple kernel learning (MKL) methods are eliminated. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1016/j.neucom.2019.01.010 | Neurocomputing |
Keywords | Field | DocType |
Hyperspectral images,Adaptive boosting,Composite kernels,Hybrid kernels,Extreme learning machines | Kernel (linear algebra),Pattern recognition,Extreme learning machine,Sparse approximation,Hyperspectral imaging,Ground truth,Artificial intelligence,Boosting (machine learning),Kernel method,Mathematics,Kernel (statistics) | Journal |
Volume | ISSN | Citations |
334 | 0925-2312 | 1 |
PageRank | References | Authors |
0.35 | 29 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Ugur Ergul | 1 | 3 | 2.74 |
| Gökhan Bilgin | 2 | 62 | 13.18 |