Optimization model based on attention mechanism for few-shot image classification

This work proposes MLAL, an optimization model based on attention for few-shot learning, which achieves exceptional 1-shot and 5-shot classification accuracy on MiniImagenet and TiredImagenet.