Bayesian neural networks for SAR feature extraction

Achotham D. Nair; Igor V. Ternovskiy

doi:10.1117/12.323149

Translator Disclaimer

2 October 1998 Bayesian neural networks for SAR feature extraction

Achotham D. Nair; Igor V. Ternovskiy

Author Affiliations +

Proceedings Volume 3464, Optical Devices and Methods for Microwave/Millimeter-Wave and Frontier Applications; (1998) https://doi.org/10.1117/12.323149
Event: SPIE's International Symposium on Optical Science, Engineering, and Instrumentation, 1998, San Diego, CA, United States

Abstract

In this paper we explore the ISMP (Isomorphic Singular Manifold Projection) technique's capability for classification. The ISMP decomposes an image into a surface sculpture or manifolds represented by polynomials and a texture image. A Bayesian neural network with Hybrid Monte Carlo learning is used to classify the texture image features. The preliminary results are encouraging and the ISMP is promising as a new texture classification scheme.

Citation Download Citation

Achotham D. Nair and Igor V. Ternovskiy "Bayesian neural networks for SAR feature extraction", Proc. SPIE 3464, Optical Devices and Methods for Microwave/Millimeter-Wave and Frontier Applications, (2 October 1998); https://doi.org/10.1117/12.323149

ACCESS THE FULL ARTICLE