A new algorithm for interlaced to progressive scan conversion based on directional correlations and its IC design

It is well known that current TV systems are suffering from uncomfortable visual artifacts such as an edge flicker, an interline flicker and a line crawling, due to the inherent nature of the interlaced scanning process. Besides, the increasing demands of large TV screens recently have turned the problem of reducing visible line structures in a TV screen into an important issue. To lessen the visual affects of those artifacts, a technique known as an interlaced to progressive conversion (IPC) technique has been widely studied in various shapes. In this paper, we propose a new algorithm for interlaced to progressive conversion, referred to as a directional correlation dependent interpolation filtering (DIF) algorithm, which is composed of a correlation-dependent linear interpolation for a signal in the low horizontal frequency band and a line doubling method for a signal lied in the high horizontal frequency band. In essence, the DIF algorithm splits the horizontal frequency band of a given input 2-Dimensional signal into two adjoint signals and apply two isolated interpolation schemes for the respective bands. In the algorithm, we develop a directional correlation dependent interpolation technique depending on the spatial correlations between the spatially symmetric samples in the observation sliding window centered at the sample to be estimated. Computer simulations and the hardware implementation of the proposed algorithm to evaluate its performance are also rigorously evaluated. To be specific, we design a chip for Luminance/Chorominance Interpolation with Peaking process (YCIP) which has been tested with a real sequence of LDP images. Simulation results validate the effectiveness of the proposed algorithm and the high performance of the YCIP in the proposed problem.