Recognition of Leaf Blight and Fruit Rot Diseases of Eggplant Using Transfer Learning

Abstract

Eggplant is an essential food in Bangladesh because the majority of the population still live below the poverty line. Eggplant is so inexpensive that it helps poor people to meet the demand of food. In other words, we can say eggplant helps to fulfill one of the basic needs (food) of humans. But a large quantity of eggplant production faces a huge loss because of eggplant’s leaf blight and fruit rot diseases. Many types of leaf blight and fruit rot diseases reduce the healthy growth of eggplant seedlings and fruits. Due to lack of technology and education in Bangladesh, many farmers are still unable to diagnose the disease properly. As a result, it is harmful for both economic development and poor people. So, we proposed our work that can detect the leaf blight and fruit rot diseases of Eggplant very accurately. We use MobileNet, CNN, Multilayer Perceptron & VGG16 for our work. MobileNet, CNN, Multilayer Perceptron, VGG16 help us to enhance the object detection result. These methods are very helpful to detect eggplant disease. MobileNet gives accuracy 95%, CNN gives accuracy 81%, Multilayer Perceptron gives accuracy 77% & VGG16 gives accuracy 70%. The accuracies are pretty good, MobileNet gives the best accuracy among the other methods. Hence this work will help to increase the economic development and eggplant production, reduce the food shortage of the poor and the grief of the farmer.