Enhanced Plant Pest Classification by Leveraging CBAM Attention in ResNet-9

This paper presents a novel approach to identifying plant pests, utilizing the Convolutional Block Attention Module (CBAM) in conjunction with the ResNet-9 architecture. The study is relevant to current issues regarding pest identification, as it aims to achieve accurate and timely rates since plant pests pose serious threats to agriculture and food security. Thus, the proposed CBAM-ResNet-9 model achieves better classification accuracy and efficiency than existing models that utilize attention mechanisms to highlight relevant image features. Transfer learning utilizes pre-trained architectures to fine-tune the model, while data augmentation helps prevent overfitting through image enhancements such as rotation, scaling, and translation. A comprehensive analysis has been conducted on publicly available pest image datasets in comparison to the performance of state-of-the-art CBAMResNet-9 models under scrutiny. The experimental results prove that the proposed model is superior to various popularly used architectures in terms of efficacy for classifying different species of pests. Such results may inspire confidence in this approach for practical implementation in agricultural contexts, enabling farmers to detect and manage pests early.