A Fruit Recognition Approach for Refrigerator Inventory Management

This study presents an image processing algorithm to identify four different classes and number of fruits. The proposed method recognizes fruits by extracting features (area, perimeter, color, eccentricity and shape) based upon the training dataset analysis. The aim is to manage refrigerator inventory by automating the process through image processing. This is done by first detecting which fruit it is and then keeping a count of it. The images have been pre-processed and used to extract features to be compared with our database. With this, an attempt has been made to avoid the hassle of last minute realization of shortage of food items. Furthermore, improvisation has been suggested based on the limitations of the proposed algorithm. It also talks about how it will be a useful aspect in the suggested application.