Denoising-based image reconstruction from pixels located at non-integer positions
This addresses a specific issue in image processing for applications requiring precise geometric transformations, but it appears incremental as it builds on existing reconstruction techniques with a denoising refinement.
The paper tackles the problem of reconstructing images from pixels at non-integer positions, which arise from operations like rotation, by proposing a method that uses triangulation-based reconstruction followed by adaptive denoising, achieving improvements of over 1.8 dB in PSNR compared to the initial estimate.
Digital images are commonly represented as regular 2D arrays, so pixels are organized in form of a matrix addressed by integers. However, there are many image processing operations, such as rotation or motion compensation, that produce pixels at non-integer positions. Typically, image reconstruction techniques cannot handle samples at non-integer positions. In this paper, we propose to use triangulation-based reconstruction as initial estimate that is later refined by a novel adaptive denoising framework. Simulations reveal that improvements of up to more than 1.8 dB (in terms of PSNR) are achieved with respect to the initial estimate.