A Hybrid Hierarchical Sparse Kernel Classification Model for Remote Sensing Image Retrieval
- https://doi.org/10.2991/ahis.k.210913.011How to use a DOI?
- Hybrid Classification, Relational Autoencoder, Relevance Vector Machine, Remote Sensing Image Retrieval, Support Vector Machine
In remote sensing applications, finding matching images across huge datasets is difficult due to the scarcity of annotated images. The high spatio-spectral resolution and high-dimensional sparse nature make the remote sensing images difficult to utilize in particular applications. Hence, competent retrieval methods are to be designed with efficient classification strategies that solve multiclass problems. This work incorporates developing a new hybrid hierarchical sparse kernel classification (HHSKC) method using relevance vector machine (RVM) and support vector machine (SVM) classifiers. The feature extraction is attained through a relational autoencoder (RAE) with proper dimensionality reduction. The excellent competing qualities of the hybrid classifiers and the deep feature representations improve the overall potential of the RAE-HHSKC framework. The proposed RAE-HHSKC is validated on two benchmark RS image datasets: the UC Merced (UCMD) dataset and the RS-19 dataset. The RAE-HHSKC framework obtained state-of-the-art results using both sparse kernel learning machines (SKLM) and deep features.
- © 2021, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - S K Sudha AU - S Aji PY - 2021 DA - 2021/09/13 TI - A Hybrid Hierarchical Sparse Kernel Classification Model for Remote Sensing Image Retrieval BT - Proceedings of the 3rd International Conference on Integrated Intelligent Computing Communication & Security (ICIIC 2021) PB - Atlantis Press SP - 77 EP - 85 SN - 2589-4900 UR - https://doi.org/10.2991/ahis.k.210913.011 DO - https://doi.org/10.2991/ahis.k.210913.011 ID - Sudha2021 ER -