Depth map estimation plays the fundamental role in 3D shape. This paper deals with one aspect of the depth map estimation, i.e., the calculation of optimum window size which is required for focus measure computation. The variation in windows size correspondingly results in variation of depth maps. Therefore, it is important to choose the correct window size for depth map computation. This paper presents a method where the data is obtained by using the same focus measure with various window sizes in the first step. In the second step, the concept of data fusion is exploited to compute the final depth map by selecting the best focus value from the data of various window sizes. © 2010 IEEE.