JohnAndrew B. Babaran, Catlyn Joy R. Bau, Diana T. Guyala, Miguel Oliver C. Pagsuyuin.
Leafguardian:forecasting leaf disease for pothos plants using convolutional neural networks
- Undergraduate Thesis: (Bachelor of Science in Computer Science) - Pamantasan ng Lungsod ng Maynila, 2024.
ABSTRACT: In the current era, the field of artificial intelligence is rapidly evolving, with a particular emphasis on its application in analyzing leaf diseases. Plant enthusiasts, especially Pothos in tropical conditions, often face challenges in identifying and treating diseases affecting these plants. The lack of knowledge about symptoms and treatments contributes to inaccuracie in maintaining the health of Pothos plants, which can lead to the spread of diseases within the species. To address these issues, a mobile application named LeafGuardian has been developed as part of research. LeafGuardian employs artificial intelligence, specifically utilizing image classification and the Convolutional Neural Network (CNN) aqlgorithm, to predict diseases in Pothos plants. The application provides insights into treatment recommendations, preventive measures, causes, and symptoms by analyzing the leaves. It can identify various diseases such as bacterial wilt, yellow leaves, and fungal leaf spots, as well as healthy leaves or no present leaves, offering guidance on prevention. The research’s evaluation incorporates a confusion matrix and ISO/IEC 25010 standards. Using a 4-Point Numerical Scale with 32 respondents selected through convenience sampling, the confusion matrix reveals high accuracy rates for identifying healthy leaves (99%), no leaves (99%), fungal leaf spot disease (93%), yellow leaf disease (91%), and bacterial wilt disease (98%). ISO/IEC 25010 assessments indicate strong agreement in functional suitability, usability, performance efficiency, and portability, with overallmean scores ranging from 3.51 to 3.64. These positive results affirm the effectiveness and user satisfaction with LeafGuardian.