The angular kernel in machine learning for hyperspectral data classification

• Published in: 2010 2nd Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing
• Main Authors: Honeine, Paul, Richard, Cédric
• Other Authors: Laboratoire Modélisation et Sûreté des Systèmes (LM2S), Institut Charles Delaunay (ICD), Université de Technologie de Troyes (UTT)-Centre National de la Recherche Scientifique (CNRS)-Université de Technologie de Troyes (UTT)-Centre National de la Recherche Scientifique (CNRS), Laboratoire Hippolyte Fizeau (FIZEAU), Université Nice Sophia Antipolis (1965 - 2019) (UNS)-Institut national des sciences de l'Univers (INSU - CNRS)-Observatoire de la Côte d'Azur, Université Côte d'Azur (UniCA)-Université Côte d'Azur (UniCA)-Centre National de la Recherche Scientifique (CNRS)
• Format: Conference Object
• Language: English
• Published: HAL CCSD 2010
• Subjects: data handling; Gaussian processes; geophysical image processing; image classification; learning (artificial intelligence); angular kernel; hyperspectral data classification; illumination energy; Gaussian kernel; urban classification task; hyperspectral images; Kernel; Support vector machines; Hyperspectral imaging; Machine learning; Spatial resolution; Hyperspectral data; spectral angle; SVM; reproducing kernel; [INFO.INFO-LG]Computer Science [cs]/Machine Learning [cs.LG]; [INFO.INFO-TS]Computer Science [cs]/Signal and Image Processing; [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing; [INFO.INFO-CV]Computer Science [cs]/Computer Vision and Pattern Recognition [cs.CV]; Iceland
• Online Access: https://hal.science/hal-01966042; https://hal.science/hal-01966042v1/document; https://hal.science/hal-01966042v1/file/10.angular.pdf; https://doi.org/10.1109/WHISPERS.2010.5594908

International audience Support vector machines have been investigated with success for hyperspectral data classification. In this paper, we propose a new kernel to measure spectral similarity, called the angular kernel. We provide some of its properties, such as its invariance to illumination energy, as well as connection to previous work. Furthermore, we show that the performance of a classifier associated to the angular kernel is comparable to the Gaussian kernel, in the sense of universality. We derive a class of kernels based on the angular kernel, and study the performance on an urban classification task.