Saliency detection based on heuristic rules

Abstract

Detecting salient regions in images aims at finding regions which contains relevant information, where a more detailed process can be applied. Saliency detection is useful in many computer vision tasks such as object segmentation, object detection, image retrieval, place recognition, among others. In this paper, we propose a method based on heuristic rules that uses color and spatial features which allows to get a good approximation to the salient region in a very low time compared with other methods in the state of the art. The tests were performed over the images of a benchmark dataset.