<![CDATA[Red chili pepper (Capsicum annuum L.) is a horticultural commodity that has high economic value, but its production is often hampered by plant disease attacks. To automatically detect diseases in chili leaves, this study uses a deep learning approach with GoogLeNet architecture and transfer learning techniques. This study aims to classify five types of chili leaf diseases, namely Healthy, Leaf Curl, Leaf Spot, Whitefly, and Yellowish, using a model initialized with pretrained weights from ImageNet. Three types of optimizers (Adam, RMSprop, and SGD) were tested to evaluate their effect on classification accuracy. The results showed that Adam performed best with a validation accuracy of 98.80%, followed by RMSprop (98.40%) and SGD (94.00%). The confusion matrix shows that misclassification occurs mainly in the Leaf Curl class, which is often confused with Yellowish, due to visual similarities. Although the classification results were excellent, limitations such as the small size of the dataset (500 images) and the need for further augmentation techniques to address prediction errors remained challenges. This research contributes to the development of an efficient and accurate computer vision-based plant disease classification system.]]>
