ABHFA-Net is a novel few-shot classification framework that models prototypes as distributions, applies spatial-channel attention, and uses Bhattacharyya-based contrastive loss, achieving state-of-the-art accuracies on benchmark and disaster datasets.
Remote Sensing 13(13), 2532 (2021)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Enhancing Few-Shot Classification of Benchmark and Disaster Imagery with ABHFA-Net
ABHFA-Net is a novel few-shot classification framework that models prototypes as distributions, applies spatial-channel attention, and uses Bhattacharyya-based contrastive loss, achieving state-of-the-art accuracies on benchmark and disaster datasets.