Improvement on Demosaicking in Plenoptic Cameras by Use of Masking Information

Abstract

Most cameras use a color filter array (CFA) to capture color images. This array has only one color among red, green and blue per pixel and so, each pixel on the sensor lacks two color channels which can be reconstructed by a process called demosaicking. This paper proposes a demosaicking method for a plenoptic camera. Plenoptic raw images have a particular lenslet structure which must be formed the empty space, unlike conventional images. These vacant spaces exist because the lens is circular. The existing demosaicking methods degrade the image quality because it interpolates missing colors, including information on the empty space. Recently, a method using information about empty spaces has been studied. However, this method uses empty space information only for G-plane. To improve this method we applied information about empty space to R/B-plane too. Experimental results show that the performance of the proposed method is superior. As compared with the existing demosaicking algorithms, we show that our solution produces the best average demosaicking performance both objectively and subjectively.