A Dual-Path Attention and Multi-Scale Fusion Network for Crop Disease and Pest Identification
Crop diseases and pests severely threaten global food security. While deep learning has shown promise in controlled settings, real-world field conditions—characterized by complex backgrounds, variable lesion scales, and high inter-class similarity—remain challenging. To overcome limitations in feature representation, s...