# Cocoa Beans drying degree assessment using machine learning _by Aubain Yro (Institut National Polytechnique Félix Houphouët-Boigny, Côte d'Ivoire) - 2019.04.11_ ###### tags: `VAADER` `Seminar` ![](https://i.imgur.com/yMQjazk.png) ## Abstract The quality of cocoa beans is one of the most important factors in ensuring its commercialization. For this, quality control of cocoa beans is a crucial step to obtain a high quality of final product. Traditionally, cocoa beans are sorted manually according to their visual characteristics for quality assurance. However, manual sorting of the cocoa beans is a time consuming procedure. It depends on a person who has been specially trained in sorting beans. This skill of the sorter varies from a person to another; therefore, it is not an accurate process. This work proposes a procedure for automatic classification of cocoa beans in order to evaluate the quality of cocoa batches. In a proposed approach, a machine vision system is designed to acquire the cocoa beans images in the lab. The beans images are processed and the targets of the beans are segmented based on k-means clustering algorithm. After that color features and texture features of segmented beans are extracted to describe the quality of the beans samples. Then, a new algorithm based on support vector machine (SVM) is introduced for the classification of the cocoa beans into classes depending on their degree of drying. Finally, experiments have been conducted on a dataset of a total of 600 images that has been used for both training and testing with 10-fold cross validation. Analysis of the results show that machine vision system coupled with SVM technique can rapidly, accurately, and reliably discriminate cocoa beans for quality assurance management. ## [Slide](https://mycore.core-cloud.net/index.php/s/EwxbNathRQM1xJs)