Food Detection with Image Processing Using Convolutional Neural Network (CNN) Method

Abstract

Currently, the payment process at restaurants is still manual and inefficient because it uses a cash register. A cashier will check what food is ordered, then count it with the cash register. This is not efficient. So food detection devices and automatic food price estimates have the answer to these deficiencies. Food detection aims to facilitate payment at restaurants, and automatic food price estimation using the Convolutional Neural Network (CNN) classification method. The detection accuracy of 6 types of food using the CNN method was obtained 100% with 80% data partition training data and 20% test data with epoch 9000 and learning rate 0.0002, with a detection time of fewer than 10 seconds.