Apple Classification and Fruits Healthy Condition Identification using YOLO-leaky v4 Architecture

Abstract

Detecting fruit can be challenging for the upcoming generations and the people living in cities since they get less exposure to the environment and detect the type of fruit is even hard like categories of apple. There are apps that classifies fruits but still none of them tells whether the fruit is healthy or not. We are trying to make a system which will not only detect the apples but also classify the category of it. There are different types of apples in the market, green apple, golden apple, red apple, dark red apple. We are trying to first detect the apple’s region ROI and then take the ROI data and pass it to the classification unit which will then classify into type of the apple also detect the health of it. This process is implemented in python using YOLO-leaky v4 architecture to detect the location of apple in the given apple image and then use Alex Net to classify it into green apple or red apple or rotten apple, which shows superior performance of proposed system.