<p>Few-shot learning aims to generate a classifier capable of identifying novel concepts with limited instances. While most existing approaches in this field focus on general object recognition, the ability to quickly identify novel concepts with subtle differences using scarce annotation is crucial for perceptual understanding in real-world scenarios. To address this issue, the paper presents a Complementary Loss Coupling for Feature-Enhanced (CLCFE) few-shot fine-grained visual recognition. In this approach, a dual-stream attention module is utilized to enhance the embedded features extracted from the query and support images. This module builds global contextual relations by incorporating structural and semantic information. Additionally, a complementary loss coupling module is designed to obtain intensified discriminative prototypes, through mining fine-grained differences among inter-class features and exploring the correlation among intra-class features. To validate the effectiveness of the proposed modules, interpretable visualizations and detailed ablations are conducted. Comparative experiments demonstrate that the CLCFE approach outperforms the previous models by a large margin in 4 few-shot fine-grained benchmarks.</p>