Identifying Fake Rice Using Computer Vision in Perspective of Bangladesh

Abstract

Bangladesh is an agricultural country. Rice is one of the major agricultural products of our country. Rice is the staple food of most of the people in the country. As a result, there is a considerable demand for rice in our country. So, there are many types of rice supply in the market. But lately it is seen that fake rice is sold in the market along with real rice. These rice cannot be easily identified with the naked eye whether it is real or fake. Now, it is often seen that people are being deceived by buying fake rice instead of real rice. Which is deadly harmful to health. For this we have collected real and fake rice and made a large dataset by capturing their images. We have preprocessed and augmented our dataset, creating two different computer vision models and checked their accuracy of identification if the rice is real or fake. Two algorithms are being used here one is Artificial Neural Network (ANN) and another one is Convolutional Neural Network (CNN). Among them CNN algorithm gives the highest accuracy which is 98%.