Intelligent classification models for food products basis on morphological, colour and texture features

The aim of this paper is to build a supervised intelligent classification model of food products such as Biscuits, Cereals, Vegetables, Edible nuts and etc., using digital images. The Correlation-based Feature Selection (CFS) algorithm and 2nd derivative pre-treatments of the Morphological, Colour and Texture features are used to train the models for classification and detection. The best prediction accuracy is obtained for the Multilayer Perceptron (MLP), Support Vector Machines (SVM), Random Forest (RF), Simple Logistic (SLOG) and Sequential Minimal Optimization (SMO) classifiers (more than 80% of the success rate for the training/test set and 80% for the validation set). The percentage of correctly classified instances is very high in these models and ranged from 80% to 96% for the training/test set and up to 95% for the validation set.

• Autores: Veernagouda Ganganagowder, Narendra; Kamath, Priya
• Tipo de recurso: Article of journal
• Fecha de publicación: 2017
• Institución: Universidad Nacional de Colombia
• Repositorio: Universidad Nacional de Colombia
• Idioma: spa
• OAI Identifier: oai:repositorio.unal.edu.co:unal/61055
• Acceso en línea: https://repositorio.unal.edu.co/handle/unal/61055; http://bdigital.unal.edu.co/59863/
• Palabra clave: 55 Ciencias de la tierra / Earth sciences and geology; 63 Agricultura y tecnologías relacionadas / Agriculture; Algorithm; digital images; food classifiers; prediction accuracy; training/test
• Rights: openAccess
• License: Atribución-NoComercial 4.0 Internacional