This paper documents an initial investigation into the effect of image degradation on the performance of transfer learning (TL) as the number of retrained layers is varied, using a well-documented, commonly-used, and well- performing deep learning classifier (VGG16). Degradations were performed on a publicly-available data set to simulate the effects of noise and varying optical resolution by electro-optical (EO/IR) imaging sensors. Performance measurements were gathered on TL performance on the base image-set as well as modified image-sets with different numbers of retrained layers, with and without data augmentation. It is shown that TL mitigates against corrupt data, and improves classifier performance with increased numbers of retrained layers. Data augmentation also improves performance. At the same time, the phenomenal performance of TL cannot overcome the lack of feature information in severely degraded images. This experiment provides a qualitative sense of when transfer learning cannot be expected to improve classification results.
Hyperspectral sensors collect data across a wide range of the electromagnetic spectrum, encoding information about the materials comprising each pixel in the scene as well as atmospheric effects and illumination conditions. Changes in scene illumination and atmospheric conditions can strongly affect the observed spectra. In the long- wave infrared, temperature variations resulting from illumination changes produce widely varying at-aperture signals and create a complex material identification problem. Machine learning techniques can use the high- dimensional spectral data to classify a diverse set of materials with high accuracy. In this study, classification techniques are investigated for a long-wave hyperspectral imager. A scene consisting of 9 different materials is imaged over an entire day providing diversity in scene illumination and surface temperatures. A Support Vector Machine classifier, feedforward neural network, and one-dimensional convolutional neural network (1D-CNN) are compared to determine which method is most robust to changes in scene illumination. The 1D-CNN outperforms the other classification methods by a wide margin when presented hyperspectral data cubes significantly different from the training data distribution. This analysis simulates real-world classifier use and validates the robustness of the 1D-CNN to changing illumination and material temperatures.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.