A Computer Vision and deep CNN Modeling for Spices Recognition

Abstract

In the countries of South Asia, there is a significant appetite for the consumption of spices. Each type of spice offers a distinctive flavor, aroma, and quality. People are unable to identify spices and do not make effective use of them, which leads to a loss of the benefits associated with certain spices and a waste of our time. Therefore, the identification of spices is necessary for the use of appropriate spices. In the context of this study, the model that we have proposed is capable of accurately detecting spices by utilizing computer vision and neural networks (CNNs) with the assistance of photographs. Within our dataset, we have 8377 photos that can be used to train our computer. We were able to achieve some fantastic results with the strategy, and despite the lack of test data and evaluation outcomes, we decided to go with the model that had the highest level of success. It achieved an accuracy of 99% using the model M1.