On semantic object detection with salient feature

Abstract

Salient detection is one of the major interests with the computer vision research. Most of the existing object-detection work focuses on the efficiency and accuracy of detecting kinds of objects without attending the saliency of the objects. In this paper, we apply salient features in a traditional object detection process with an algorithm combining both bottom-up and top-down approaches, aiming to detect meaningful objects exhibiting saliency. We define salient region as distinct areas detected in an image in a bottom-up phase, and the salient feature as semantic features in representing an object in the top-down phase where we apply a boosting algorithm to accommodate kinds of classifiers including Support Vector Machine (SVM) and an enhanced Adaboost classifiers. Final experiments indicate that our proposed object detector is fairly effective compared with the state of the art. © 2008 Springer Berlin Heidelberg.