An intelligent cocoa quality testing framework based on deep learning techniques
Richard Essah, Darpan Anand, Surender Singh
Abstract
The export of cocoa beans is economically significant in Ghana and numerous other tropical nations. Before the cocoa beans can be processed into finished goods, they must undergo a post-harvest treatment. Fermentation is one of the most crucial steps in this process because it improves the final product's quality. Traditionally, a cut-test is used to evaluate the degree of cocoa fermentation for quality control purposes. Nonetheless, this strategy is subjective and has a number of limitations. In this paper, an attempt has been made to propose a quality testing framework for cocoa beans using deep learning techniques. There are various deep learning techniques that are used by various authors, such as CNN and DNN. Here, we use the transfer learning method to check the quality of cocoa beans. The proposed mechanism is implemented using Python in the Jupiter Notebook.