An Efficient Approach For To Predict The Quality Of Apple Through Its Appearance

Freshness attracts buyers and money for sellers. How effective will it be to automatically classify vegetables and fruits according to the freshness rate? Well, the next question comes from each one of us, is it possible or not? Yes, it is possible. This study focuses on the classification of apples based on parameters for setting their freshness. For the sake of classification, we have trained the model using CNN i.e., convolutional neural networks. Using this model apples are classified based on their appearance which is one of the most common approaches taken by humans while manually classifying apples. It is an easy, time-effective, and precise method that we prefer as the basic or most popular parameter to define the freshness of any fruit and vegetable is through their image. Using this approach, we have used CNN architectures VGG16 and Inception V3 and their accuracies were 86% and 93%. The model which is being created by us is tested on four activation functions such as read, sigmoid, tanh, and linear.