Comparison of novelty detection methods for multispectral images in rover-based planetary exploration missions
- PDF / 3,518,553 Bytes
- 34 Pages / 439.37 x 666.142 pts Page_size
- 46 Downloads / 182 Views
Comparison of novelty detection methods for multispectral images in rover-based planetary exploration missions Hannah R. Kerner1 · Kiri L. Wagstaff2 · Brian D. Bue2 · Danika F. Wellington1 · Samantha Jacob1 · Paul Horton1 · James F. Bell III1 · Chiman Kwan3 · Heni Ben Amor1 Received: 2 September 2019 / Accepted: 3 June 2020 © The Author(s) 2020
Abstract Science teams for rover-based planetary exploration missions like the Mars Science Laboratory Curiosity rover have limited time for analyzing new data before making decisions about follow-up observations. There is a need for systems that can rapidly and intelligently extract information from planetary instrument datasets and focus attention on the most promising or novel observations. Several novelty detection methods have been explored in prior work for three-channel color images and nonimage datasets, but few have considered multispectral or hyperspectral image datasets for the purpose of scientific discovery. We compared the performance of four novelty detection methods—Reed Xiaoli (RX) detectors, principal component analysis (PCA), autoencoders, and generative adversarial networks (GANs)—and the ability of each method to provide explanatory visualizations to help scientists understand and trust predictions made by the system. We show that pixel-wise RX and autoencoders trained with structural similarity (SSIM) loss can detect morphological novelties that are not detected by PCA, GANs, and mean squared error autoencoders, but that the latter methods are better suited for detecting spectral novelties—i.e., the best method for a given setting depends on the type of novelties that are sought. Additionally, we find that autoencoders provide the most useful explanatory visualizations for enabling users to understand and trust model detections, and that existing GAN approaches to novelty detection may be limited in this respect.
Responsible editor: Indre Zliobaite.
B
Hannah R. Kerner [email protected]
1
Arizona State University, 781 E Terrace Mall, Tempe, AZ 85287, USA
2
Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA
3
Applied Research, LLC, 9605 Medical Center Drive, Suite 113E, Rockville, MD 20850, USA
123
H. R. Kerner et al.
Keywords Novelty detection · Unsupervised learning · Space exploration
1 Introduction The goal of novelty detection approaches is to identify patterns in data that have not been previously observed (Markou and Singh 2003a, b; Chandola et al. 2009; Pimentel et al. 2014). The exact definition of “novelty” varies depending on the application domain and the type of data, but in all cases novel examples differ in some way from “normal” data (Pimentel et al. 2014) and are of particular interest to the user (Chandola et al. 2009). In many real-world applications, novelty detection can provide significant, actionable information, such as a novel feature in a medical image may indicate the presence of a disease or tumor (Schlegl et al. 2017), or novelty in an X-ray scan at airport
Data Loading...
