DESIGN A MACHINE LEARNING MODEL FOR SEGMENTATION OF IRREGULAR SHAPE FRUIT IMAGE CAPTURED IN NATURAL LIGHT TO IDENTIFY ARTIFICIALLY RIPENED FRUIT
Abstract :The segmentation of an image of an irregularly shaped fruit that was taken in natural light is a new way that visual recognition and machine learning can be used to find artificially ripened fruit. When eaten by people, ripe fruits can spread a wide range of infectious diseases. Several different classifiers were used to look at the fruits’ colour (measured in the RGB colour space), shape, and texture in order to find out which one was the most accurate. After the first step of picture segmentation using the FCM Based Enhanced Technique, the suggested method uses a modified Convolutional Neural Network to sort the images into one of several groups. The study’s authors suggest using a convolutional neural network (CNN) to make predictions about mangos that have been ripened in a lab. By doing this, they hope to cut down on the amount of money that will be wasted because of their work. Using this method, mango fruits that have been ripened in a lab can be put into the right category more accurately and with less waste. The artificially ripened mango fruit method of predicting the future shows that things are going to get better soon. Based on the results of our tests, the proposed method can help a lot with accurate identification and automatic classification of fruits with odd shapes in natural light images. This can be done to tell the difference between fruits that have ripened on their own and those that have been made to ripen. When figuring out what characteristics ripe fruit has, a grey-level co-occurrence matrix is used. The fruits that are being looked at right now are apples, oranges, grapes, pomegranates, and bananas.
Keywords: Irregular Shape Fruit Image, convolutional neural network, FCM.