<![CDATA[Apple plant classification based on leaf images still faces challenges in the form of variations in shape, texture, lighting, and visual similarity with other plants, which can potentially reduce identification accuracy. This problem impacts agricultural monitoring and decision-making processes that require high accuracy. This research aims to apply the Convolutional Neural Network (CNN) method, a deep learning approach designed for automatically extracting visual features, to apple plant classification using a fifty-layer Residual Network (ResNet-50) architecture. ResNet-50 uses a residual mechanism to overcome performance degradation in very deep networks. The results showed that the CNN model based on the ResNet-50 architecture was able to classify healthy and rotten apple leaf images with the highest accuracy of 82.40% using the 80:20 dataset split scenario. Evaluation using a confusion matrix and classification report produced a precision value of 0.86 for the rotten class and 0.79 for the healthy class, while the recal values were 0.77 and 0.88, respectivelty. These rindings indicate that ResNet-50 has a good capability in recognizing visual characteristics of apple leaves and can be effectively utilized as an image-based plant classification solution.]]>
Copyrights © 2026
