Bayesian neural networks to analyze hyperspectral datasets using uncertainty metrics

Alcolea, Adrián (Universidad de Zaragoza) ; Resano, Javier (Universidad de Zaragoza)

Resumen: Machine learning techniques, and specifically neural networks, have proved to be very useful tools for image classification tasks. Nevertheless, measuring the reliability of these networks and calibrating them accurately are very complex. This is even more complex in a field like hyperspectral imaging, where labeled data are scarce and difficult to generate. Bayesian neural networks (BNNs) allow to obtain uncertainty metrics related to the data processed (aleatoric), and to the uncertainty generated by the model selected (epistemic). On this work, we will demonstrate the utility of BNNs by analyzing the uncertainty metrics obtained by a BNN over five of the most used hyperspectral images datasets. In addition, we will illustrate how these metrics can be used for several practical applications such as identifying predictions that do not reach the required level of accuracy, detecting mislabeling in the dataset, or identifying when the predictions are affected by the increase of the level of noise in the input data.

Idioma: Inglés

DOI: 10.1109/TGRS.2022.3205119

Año: 2022

Publicado en: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 60 (2022), 5537810 [10 pp.]
ISSN: 0196-2892

Factor impacto JCR: 8.2 (2022)

Categ. JCR: ENGINEERING, ELECTRICAL & ELECTRONIC rank: 27 / 274 = 0.099 (2022) - Q1 - T1
Categ. JCR: IMAGING SCIENCE & PHOTOGRAPHIC TECHNOLOGY rank: 5 / 28 = 0.179 (2022) - Q1 - T1
Categ. JCR: REMOTE SENSING rank: 4 / 34 = 0.118 (2022) - Q1 - T1
Categ. JCR: GEOCHEMISTRY & GEOPHYSICS rank: 5 / 87 = 0.057 (2022) - Q1 - T1

Factor impacto CITESCORE: 10.9 - Earth and Planetary Sciences (Q1) - Engineering (Q1)

Factor impacto SCIMAGO: 2.4 - Electrical and Electronic Engineering (Q1) - Earth and Planetary Sciences (miscellaneous) (Q1)

Financiación: info:eu-repo/grantAgreement/ES/AEI/PID2019-105660RB-C21-AEI-10.13039-501100011033
Financiación: info:eu-repo/grantAgreement/ES/DGA-ESF/T58-20R

Tipo y forma: Artículo (PostPrint)

Área (Departamento): Área Arquit.Tecnología Comput. (Dpto. Informát.Ingenie.Sistms.)