Application of Deep Learning for Mobile-Based Detection of Leaf Diseases in Amorphophallus Muelleri Blume Using SSD-Mobilenet

This study aims to develop a mobile-based deep learning system for precise Amorphophallus muelleri Blume leaf disease detection and classification. The diseases, primarily caused by fungi, nematodes, and larvae, often exhibit similar leaf symptoms. By implementing the SSD-MobileNet method, our proposed system achieves the highest accuracy rate of 95.45% with the farthest distance that can be detected by the system is 15 cm in real-time classification. Additionally, through image resizing to 640x640 pixels and fine-tuning parameters, including epochs, momentum, and learning rates, our approach demonstrates high-performance suitability for mobile platforms.