Generation of Depth Map Based on Depth from Focus: A Survey

The important step in 3D data generation for 3D display system is the estimation of depth map. Depth map is a gray-level image which has exactly the same size of the original captured 2D image data that indicates the relative or apparent distance of each pixel from the camera to the objects in the real world. The Depth From Focus techniques use a search for the most in-focus position over a sequence of images taken by varying a camera parameter i.e. "focus" little by little, therefore high spatial and temporal resolution cannot be obtained because of an output bottleneck of an image sensor. This paper presents a survey of some practical approaches which are already implemented to generate depth map from a single image or series of images taken with different focuses of the same scene. Smoothing filters like 2D Gaussian Filter, Asymmetric Filter with some addition or subtraction of functional blocks can be used which are easy to implement, and the result is robust to any content of scene. These filters are applied on synthesized depth map for preprocessing to reduce blocky artifacts. Various methods from image capturing by any focusing camera with multiple focus values to the depth map generation and the final 3D data rendering is discussed here.