In this paper we present an algorithm of depth estimation from a single frame captured with color coded aperture camera. Our algorithm provides continuous depth and is more robust to lack of texture comparing to the state-of-the-art. The main contributions of this work comparing to prior-art algorithms are: (1) robust metric for depth determination in a pixel, (2) depth map sub-pixel estimation, (3) depth propagation to low-textured areas, (4) depth map edge restoration, (5) depth quality enhancement, (6) raw data processing. We also made an efficient algorithm implementation which processes FullHD frame for 50 ms on GeForce GTX 780 and requires 15 seconds on Qualcomm Andreno 330 GPGPU for the same frame.
Ivan Panchenko, Vladimir Paramonov, Victor Bucha, "Depth Estimation Algorithm for Color Coded Aperture Camera" in Proc. IS&T Int’l. Symp. on Electronic Imaging: 3D Image Processing, Measurement (3DIPM), and Applications, 2016, https://doi.org/10.2352/ISSN.2470-1173.2016.21.3DIPM-405