DeepSeqCoco: A Robust Mobile Friendly Deep Learning Model for Detection of Diseases in Cocos Nucifera

Cocos Nucifera, commonly known as the coconut tree, plays a vital role in tropical agriculture and rural livelihoods. However, the productivity and sustainability of coconut cultivation are increasingly threatened by various fungal and bacterial diseases, including Bud Root Dropping, Bud Rot, Gray Leaf Spot, Leaf Rot, and Stem Bleeding. Manual identification of these diseases is time-consuming, error-prone, and requires expert knowledge, making automated approaches highly desirable. This study introduces DeepSeqCoco, a robust and mobile-friendly deep learning framework for early detection and classification of coconut diseases using image processing and transfer learning. The proposed model leverages the EfficientNet-B3 architecture, optimized for high accuracy and computational efficiency, enabling deployment on mobile and resource-constrained devices. A comprehensive dataset of coconut leaves and stem images was preprocessed with resizing, augmentation, and normalization to improve generalization and reduce overfitting. The model achieved high accuracy, precision, recall, and F1-score, as validated through confusion matrix analysis. Experimental results confirm that DeepSeqCoco can reliably distinguish between the five major coconut diseases, thus empowering farmers with a practical and automated disease detection tool. The framework demonstrates the potential of deep learning for sustainable agriculture and provides a scalable pathway for mobile-based crop health monitoring.

Index Terms— DeepSeqCoco, Disease Detection, Cocos Nucifera, EfficientNet-B3, Image Processing.