Analyzing Image Focus using Deep Neural Network for 3D Shape Recovery

Shape from Focus is one of the passive optical techniques for 3D shape recovery, from a set of differently focused 2D images. It utilizes the focus information present in the image to find 3D shape of the object in consideration. First, a stack of images is obtained by moving the object along the optical axis. Then a Focus Measure operator is applied leading to the focus curves which are then maximized to obtain the best focused positions. Conventionally, the previously proposed focus measures are computationally expensive since they have to process huge amounts of data. In this paper, Deep Neural Networks (DNN) have been employed to measure the amount of focus in the image stack. The results are compared with commonly used FMs by employing RMSE, Correlation and $Q$ index. The comparison establishes that the proposed method is not only only efficient but more accurate.