Image segmentation is an essential step in image analysis.A novel model was proposed for unsupervised segmentation of viewers attention object from natural images by using salient object edge points and localizing regionbased active model (LRAC).The salient edge points were firstly detected by the color boosting Harris detector.The core saliency map was then proposed to obtain the foreground seeds from the salient edge points.Thenthe salient object edge points were determined by the foreground seeds.Finallythe salient object edge points were employed as the seeds of convex hull to generate the initial contour for our automatic object segmentation system.Extensive experiments results on a large variety of natural images prove that the proposed model can provide satisfactory segmentation performance.