MSFNet-CPD: Multi-Scale Cross-Modal Fusion Network for Crop Pest Detection

MSFNet-CPD is a multi-scale cross-modal fusion network for crop pest detection, addressing challenges like intra-class variance and limited datasets. It introduces a super-resolution module, an Image-Text Fusion module, and an Arbitrary Combination Image Enhancement strategy to improve feature modeling and model generalization. MSFNet-CPD outperforms state-of-the-art methods on multiple pest detection benchmarks.