Texture- and Symptom-Based Coffee Leaf Disease Classification Using CNN and ResNet algorithms

This work aims to improve plant disease prediction through the implementation of an intriguing and revolutionary model that combines refined algorithms for machine learning with predictive analytics strategies. The goal is to strengthen utilities' resilience by providing precise and trustworthy forecasts of the potential classification of diseases. This study aims to push the limits of present prediction skills, analyze cutting-edge technology, and test unique features to improve the accuracy and timeliness of plant pathogen prophecies significantly. The ultimate goal in a rapidly evolving technological environment is to contribute to the development of proactive and effective mitigation techniques that reduce the impact of electrical glitches on equipment, infrastructure, and system-wide consistency. Materials required: The core database for the present investigation was obtained from the reputable data-gathering and analytics website Kaggle. The dataset includes a wide range of line voltages and currents for various failure scenarios. In Group 1, the convolutional neural network approach was applied. The statistical power of the 48 sample sizes utilized in the study was 0.8, the alpha value was 0.05, and the beta value was 0.2. ResNet is used by Group 2. Clinical is the method used to compute these figures. To run the accurate prediction software, Google Colab is employed. It analyzes data and displays accuracy as a percentage and in a graphical format for simple interpretation. Result: The convolutional neural network technique yielded an accuracy of 78.65% percent for Group 1, while the ResNet strategy yielded an accuracy of 70.61% for Group 2. SPSS data analysis program is employed for data set statistical analysis and correctness. A two-tailed test is used to appraise to what extent there is a statistically significant discrepancy in accuracy between both sets of data. Conclusion: Convolutional neural networks are more accurate than ResNet algorithms.