Image Principal Part Augmentation for Fine-Grained Image Classification

To avoid the dramatic impact due to the agricultural pests, computer vision technology used to recognize agricultural pests has been a focal spot. Nevertheless, there have been some significant challenges about fine-grained image classification. This paper proposes an image principal part augmentation method tailored for insect scenarios, aiming to address the challenges posed by subtle features and background interference in fine-grained image classification. The approach not only significantly enhances network performance but also demonstrates high versatility. This paper makes meaningful strides in methodology for fine-grained image analysis, particularly in insect scenes.