Segmentation with low depth of field images

In this paper, we propose a segmentation algorithm for separating focused foreground region from defocused background region with low depth of field(DOF) image. Low DOF is a one of the most popular photography technique for emphasize the importance object. and highly useful in various application including image enhancement for digital cameras, target recognition, image indexing for content-based retrieval, and 3D microscopic image analysis. This algorithm consists of three stages. At the first stage, we generate a blur parameter map which represent the degree of defocus value from an input image by the point spread function. Then, these parameter values are refined defocused regions and focused regions using efficient belief propagation algorithm. Experimental results show that this approach can also separate the focused region from defocused region effectively and accurately.