The Hybrid Approach of Image Segmentation Using MeanShift and Saliency Maps

Image segmentation is an important problems in the field of image processing. In this paper, we propose a hybrid approach of image segmentation, combining MeanShift, an image segmentation method, and Saliency Maps, a regions-of-interest detection method. Our proposed method uses low-level features of color image, such as luminance, color and spatial. The results from experiments showed that our proposed method solves two problems, one is determining the regions of objects with high saliency in image and the other is associating discrete broken regions based on the relationship between low-level features and saliency feature.