Feature Channel Adaptive Enhancement for Fine-Grained Visual Classification

Fine-grained classification poses greater challenges comparedto basic-level image classification due to the visually similar sub-species.To distinguish between confusing species, we introduce a novel frameworkbased on feature channel adaptive enhancement and attention erasure.On one hand, a lightweight module employing both channel attentionand spatial attention is designed, adaptively enhancing the feature expression of important areas and obtaining more discriminative featurevectors. On the other hand, we incorporate attention erasure methodsthat compel the network to concentrate on less prominent areas, therebyenhancing the network’s robustness. Our method can be seamlessly integrated into various backbone networks. Finally, an evaluation of ourapproach is conducted across diverse public datasets, accompanied bya comprehensive comparative analysis against state-of-the-art methodologies. The experimental findings substantiate the efficacy and viabilityof our method in real-world scenarios, exemplifying noteworthy breakthroughs in intricate fine-grained classification endeavors.