An Inpainting- Based Deinterlacing Method

Abstract

Abstract-Video is usually acquired in interlaced format, where each image frame is composed of two image fields, each field holding same parity lines. However, many display devices require progressive video as input; also, many video processing tasks perform better on progressive material than on interlaced video. In the literature, there exist a great number of algorithms for interlaced to progressive video conversion, with a great tradeoff between the speed and quality of the results. The best algorithms in terms of image quality require motion compensation; hence, they are computationally very intensive. In this paper, we propose a novel deinterlacing algorithm based on ideas from the image inpainting arena. We view the lines to interpolate as gaps that we need to inpaint. Numerically, this is implemented using a dynamic programming procedure, which ensures a complexity of ( ), where is the number of pixels in the image. The results obtained with our algorithm compare favorably, in terms of image quality, with state-of-the-art methods, but at a lower computational cost, since we do not need to perform motion field estimation.

3-D Data Denoising and Inpainting with the Low-Redundancy Fast Curvelet Transform

“…We will be left to a future work. Figures 12, 13, 14, 15 show close-ups of the de-interlacing results on the test sequences stockholm and shields, and compare them the methods proposed in [32,37,33]. We will refer to them respectively as DOI, STIM, and STELA.…”

“…The problem of reconstructing the missing lines in the sequence is called de-interlacing and has been addressed with many dedicated methods : using simple spatio-temporal neighborhood linear interpolation [29,30], using a directional interpolation to better preserve the edges in the images [31,32,33], with more complicated methods relying on motion estimation and compensation [34,35,36], and recently with variational inpainting [37]. In this section, it will be demonstrated that adpating the idea of sparsity-based inpainting as described in Section 3.2, and introducing a proper regularization that takes into account the special structure of the mask, will yield very promising de-interlacing results.…”

“…We also wish to thank Luis Garrido for giving us the de-interlaced sequences presented in [37] to which we compare in Section 3.3.…”

Geometry-Based Demosaicking

“…The algorithms using hue-based interpolation first interpolate I G and then assume perfect correlation among channels to interpolate I R and I B , see [2]. The same hypothesis is used by state of the art methods in other color interpolation problems such as colorization of grayscale images [16], [17], [18], deinterlacing [19] or movie denoising [20].…”

“…1 (c)) and the optimization problem can be solved with dynamic programming as Cox et al propose in the context of image stereo in [23]. See [19] for an application of this same approach to de-interlacing. The use of dynamic programming for inpainting was proposed by Masnou in [10].…”