Extreme Value Meta-Learning for Few-Shot Open-Set Recognition of Hyperspectral Images
Debabrata Pal, Shirsha Bose, Biplab Banerjee, Yogananda Jeppu
Abstract
Recent advancements in prototype-based Few-Shot Open-Set Recognition (FSOSR) approaches reject outliers based on the high metric distances from the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known</i> class prototypes and fail to distinguish spectrally fine-grained land cover outliers. Learning only the Euclidean distance fit spherical distributions ignores the essential distribution parameters like shift, shape, and scale. The conventional meta-training of FSOSR also ignores the topological consistency of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known</i> classes impacting reduced closed and open accuracy in the meta-testing phase. Moreover, the existing hyperspectral outlier detection methods do not provide intuition about the rejected outlier’s land cover category. To tackle the aforesaid problems, we introduce <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Extreme Value Meta-Learning</i> (EVML), where we fit Weibull distributions per known class based on the limited support-set distances from respective prototypes. A newly proposed Prototypical OpenMax (P-OpenMax) layer leverages these meta-trained Weibull models and calibrates the query distances to reject fine-grained outliers. Then, to learn the topological consistency, we split all the samples in an episode into four parts, including the prototype and its same <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known</i> class queries, other <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known</i> class queries, and the remaining <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known-unknown</i> queries. A novel open quadruplet loss ensures that a prototype’s same-class queries reside closer than the other <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known</i> -class and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">known-unknown</i> queries. Finally, we coarse classify the detected outliers into major land cover categories and perform cross-dataset incremental FSOSR to enhance robustness over unknown geographical regions. We validate the efficacy of EVML over four benchmark hyperspectral datasets.