A. Elmoataz et al. (Eds.): ICISP 2008, LNCS 5099, pp. 120 – 127, 2008. © Springer-Verlag Berlin Heidelberg 2008 Depth Estimation by Finding Best Focused Points Using Line Fitting Aamir Saeed Malik and Tae-Sun Choi Gwangju Institute of Science and Technology, 1 Oryong-Dong, Buk-Gu, Gwangju, 500712, Korea {aamir, tschoi}@gist.ac.kr Abstract. This paper presents a method for depth estimation using image focus based on the linear regression model. For each pixel, we select two datasets based on the maximum value which is calculated using Laplacian operator. Then linear regression model is used to find lines that approximate these data- sets. The best fit lines are found using least squares method. After approximat- ing the two lines, their intersection point is calculated and weights are assigned to calculate the new value for the depth map. The proposed method is compared with four depth estimation algorithms. Keywords: Depth Map, 3D shape recovery, Shape from focus, Line fitting. 1 Introduction There are many methods for the calculation of depth leading to 3D shape recovery. In this paper, we limit our discussion to one of such methods, i.e., Shape From Focus (SFF). The objective of shape from focus is to find out the depth of every point of the object from the camera lens. Hence, finally we get a depth map which contains the depth of all points of the object from the camera lens where they are best focused or in other words, where they show maximum sharpness. The basic problem of imaging systems, such as the eye or a video-camera, is that depth information is lost while projecting a 3D scene onto 2D image plane. Therefore, one fundamental problem in computer vision is the reconstruction of a geometric object from one or several observations. Shape information that is obtained from the reconstruction of a geometric object is of critical importance in many higher level vision applications like mobile robot systems. For example, an unmanned spacecraft, in order to land safely on lunar surface, needs to estimate depth details of the terrain. Various image processing techniques retrieve the lost cue and shape information from the pictorial information. Shape from focus (SFF) is one of such image processing techniques that are used to recover such information. Various techniques and algorithms have been proposed in the literature for the imple- mentation of SFF. They include methods using focus image surface, Lagrange polyno- mial, neural networks, dynamic programming etc. But almost all the techniques start with the estimation of the depth map. Hence, the techniques for the estimation of this initial depth map become quite significant.