Regresar

Table Olive Classifier Model for the Preparation of Gourmet Dishes

Abstract:

The consumption of olives has increased due to their use in the field of gastronomy. The success of olive production originates in European countries and extends to Latin American countries such as Peru, Argentina, and Uruguay. There is a wide variety of olives, among which table olives are the focus of this research due to their quality and level of ripeness at the time of harvesting. The objective of this work is to develop a classifier for table olive varieties that considers shape, color, and texture characteristics to offer better quality in the preparation of gourmet dishes. The development of this work utilizes open-source programs and libraries for Machine Learning (ML), including Python, TensorFlow, Keras, and the YOLO (You Only Look Once) algorithm, which has become very popular for object detection. The dataset consists of 3000 images, and the methodology follows the Digital Image Processing (DIP) approach, which consists of five phases: capture, preprocessing, segmentation, feature extraction, and classification. A classifier based on CNN (Convolutional Neural Network) has been selected, and the results achieved an accuracy of 98% for the Manzanilla olive and 97% for the Botija olive.