Detection of Cucumber Diseases using Deep Learning models

Cucumber extensively grown vegetable, is highly regarded for its high-water content, low calorie value, and have many health benefits, such as its ability to reduce blood sugar, fight bacteria, and boost antioxidant levels. However, a variety of environmental factors can cause cucumbers to become susceptible to a wide range of diseases, such as powdery mildew, downy mildew, and bacterial wilt which can have a significant impact on both yield and quality. An accurate and reliable automated diseases detection system must be developed for identification of cucumber diseases. We examined the performance of many deep models and assessed their efficacy in diagnosing the cucumber diseases in this study by using deep learning models with data augmentation. We compared the performance of AlexNet, ResNet50, and GoogLeNet and analyzed their efficiency in terms of model accuracy, F1 score and confusion matrix in detection of cucumber diseases images. In this study, we compared the accuracy and F1 score of AlexNet, GoogleNet, and ResNet50 models. We observed that the accuracy of AlexNet, GoogleNet, and ResNet50 models is 0.85, 0.87, and 0.90, and their F1 scores are 0.76, 0.79, and.089, respectively.