A Progressive Feature Learning Network for Cordyceps sinensis Image Recognition

<i>Cordyceps sinensis</i> (<i>C. sinensis</i>) is a valuable herbal medicine with wide-ranging applications. However, automating <i>C. sinensis</i> recognition is challenging due to the high morphological similarity and limited phenotypic variation among its subspecies. In this paper, we propose a novel approach called Progressive Feature Learning Network (PFL-Net) that mines multiple biological features to recognize different subspecies. Firstly, to comprehensively capture multi-scale discriminative features of <i>C. sinensis</i>, we propose the Spatial-aware Semantic Refinement Module (SSRM), which constructs discriminative feature groups by utilizing relative positions to model the intrinsic feature relations. Secondly, the Multi-scale Collaborative Perception Module (MCPM) avoids isolated biological features during modeling by establishing relations between different feature groups to enhance the recognition integrity of <i>C. sinensis</i>. Furthermore, to prevent the model from focusing on the same discriminative regions of <i>C. sinensis</i>, we propose a Channel Decouple (CD) loss that decouples features along the channel dimension, enhancing the diversity of <i>C. sinensis</i> discriminative features. In addition, we construct a <i>C. sinensis</i> dataset (CSD) to facilitate the application of biometric recognition, representing the first study focused on fine-grained <i>C. sinensis</i> recognition. Extensive experiments conducted on the CSD and three benchmark datasets validate the effectiveness of our proposed method, achieving a top-1 accuracy of 94.43% on the CSD dataset, which surpasses all existing approaches.