A Deep Learning-Based Crop Disease Diagnosis Method Using Multimodal Mixup Augmentation

• DOI: 10.3390/app14104322
• Publication Date: 2024-05-20T12:20:31Z
• Authors: Hyunseok Lee; Young-Sang Park; Songho Yang; Hoyul Lee; Tae-Jin Park; Doyeob Yeo
• Keywords: Technology; diagnosis; QH301-705.5; T; Physics; QC1-999; multimodal; deep learning; 02 engineering and technology; mixup augmentation; Engineering (General). Civil engineering (General); Chemistry; 0202 electrical engineering, electronic engineering, information engineering; TA1-2040; Biology (General); QD1-999; crop disease

With the widespread adoption of smart farms and continuous advancements in IoT (Internet Things) technology, acquiring diverse additional data has become increasingly convenient. Consequently, studies relevant to deep learning models that leverage multimodal for crop disease diagnosis associated augmentation methods are significantly growing. We propose a comprehensive model predicts type, detects presence, assesses severity at same time. utilize comprising images environmental variables such as temperature, humidity, dew points. confirmed results diagnosing diseases using improved 2.58%p performance compared only. also multimodal-based mixup method capable utilizing both image data. In this study, refer from multiple sources, is technique combines training. This expands conventional was originally applied solely Our showcases improvement 1.33%p original method.