Cassava Disease Recognition and Model Analysis Based on Joint ResNeXt and EfficientNet

Cassava, cultivated by smallholder farmers, is a crop essential to food security, but their yields are often low due to the crop's vulnerability to viral infections. Given that farmers use low-bandwidth, mobile-quality cameras, an effective technique for detecting illness under severe restrictions is desperately needed. This paper proposes a merged model of ResNeXt and EfficientNet to construct a recognition model. In order to provide a more thorough understanding of the diverse and complex nature of cassava leaf diseases, this model aims to classify each cassava image into four disease categories or a fifth category indicating a healthy leaf. It also identifies the specific differences between different health states of cassava leaves. This holistic approach to cassava disease recognition demonstrates the potential ability to integrate different neural network architectures, further deepening the understanding of the subtleties of cassava leaf diseases. Thus, this study makes a significant contribution to the field of cassava leaf disease recognition, demonstrating the effectiveness of integrating multi-scale feature extraction techniques to improve model performance. It establishes the groundwork for new approaches in the field of study and enables farmers to swiftly spot ill plants, maybe rescuing their harvests before they do irreversible harm.