The Rose's Mysteries: An Extensive Examination of Rose Leaf Illnesses
Main Article Content
Abstract
Rose plants, cherished for their vibrant blooms and fragrant essence, are susceptible to various diseases that can severely impact their health and aesthetic appeal. This research paper provides a comprehensive analysis of common rose leaf diseases, focusing on their causes, symptoms, and management strategies. Through a combination of literature review, field observations, and expert insights, this study aims to enhance our understanding of rose leaf diseases and equip gardeners, horticulturists, and researchers with effective tools for disease prevention and control. The findings highlight the importance of early detection, proper sanitation practices, and integrated pest management techniques in maintaining the vitality and beauty of rose plants. Accuracy without transfer learning (87.82%) significantly outperforms that with transfer learning (64.53%). The highest accuracy was achieved without transfer learning (95.00%), compared to 81.56% with transfer learning. Without transfer learning, the model achieves 87.15% accuracy, while with transfer learning, it is limited to 67.28%. Accuracy without transfer learning (94.37%) is markedly better than with transfer learning (72.88%). The characteristic rust-colored spots on leaves are better detected with a specialized model. Across all four diseases, models trained without transfer learning consistently outperform those utilizing transfer learning. This emphasizes the importance of disease-specific training for accurate detection. While transfer learning is often advantageous in scenarios with limited data, its effectiveness diminishes when pre-trained features fail to generalize to domain-specific tasks, such as rose disease detection.