Image Classification Using Statistical Learning for Automatic Archiving System
Jassem Mtimet, Hamid Amiri · International Review on Computers and Software (IRECOS) · 2013
Recently the volume of digital images has grown too rapidly that it is obvious that building an efficient mechanism for managing such data in a digital archive system becomes a necessary task. In this paper, we propose an image classification tool as an important module in a dedicated archiving system. This tool can be used to verify image categories (photo, textual or mixed image). The proposed technique extracts a set of low-level feature from the processed image. Two classifiers (Decision Tree and Neuronal Network) are then used to train and classify images. Our results prove that the proposed classification tools can be efficiently used to build our archiving system, with a distinct performance for each classifier, depending on the image’s type.