An SVM based system for automatic dates sorting

Abstract

In this work, we present an automatic sorting dates system based on a statistical learning technique called Support vector machine. This system is developed to be used in factories of the region of Biskra working on dates sorting and packaging. Our system analyses dates images acquired by a camera and builds a features database. The SVM technique uses then this database to search hyperplans which separate better different dates’ classes in features space. Hyperplans are used to classify new dates images. Obtained results show SVM power in dates’ images classification and validate proposed system against similar sorting systems.

Link http://www.praiseworthyprize.com/IRECOS-latest/IRECOS_vol_5_n_4.html#An_SVM_Based_System_for_Automatic_Dates_Sorting