Support Vector Machine for the Classification of Images Captured by WMSN
Astrit Hulaj, Adrian Shehu, Xhevahir Bajrami · 2017
Today with the aim of increasing cross-border security, especially along the borderline. In certain countries they have designed a system that includes an application of multimedia sensor nodes. The images captured by the multimedia sensor nodes are located along the borderline, it may occur that the images do not have important information for border security. In this paper we will present a new algorithm that will enable the classification and identification of information that contains these images. In other words, this algorithm, enables automatic classification of the image captured by the sensor node and illustrates in the output (monitor), projecting a distinct image, where it could happen to be an animal, bird or even a human being. The working principle of this algorithm is based on the logic of functioning Support Vector Machine (SVM). The results that will be presented in this paper will indicate that this algorithm, is very efficient for the classification and identification of the images captured by the multimedia sensor nodes and alerting of security authorities.