International Research Journal of Education and Technology

Nowadays we must be more concerned about the quality of the food items we consume and feed our loved ones. To get a full nutrition & mineral value in food items, it’s better to consume the fresh food items to avoid unnecessary health problems. We have various methods to determine or detect the freshness quality of our food items like the traditional method of lab testing the food sample, using images of the food sample with various machine learning and deep learning algorithms and even using the smell from the food sample. Many machine learning models have been developed to detect the freshness of food samples, but they lack fast and accurate detection. We have developed an IOT model that can predict the freshness of the food samples with image and its smell in a more accurate level with less time using deep learning algorithms. We have combined two deep learning algorithms for image detection to obtain more accurate results. We classify the images of food samples using the Convolutional Neural Network and detect the affected areas using the Object Detection Algorithm YOLO. We also take the level of gas emitted by rotten food samples using sensors and detect the spoilage level using the Artificial Neural Network model. We can determine the affected or rotten areas by combining the value from both image data and gas emitted by the given food sample. The result can be viewed in an application through a browser.

Keywords: deep learning algorithms, YOLO, CNN, ANN, IOT.