An Efficient Pre-processing Technique using HOLF for 2D Image Interpolation
Jagyanseni Panda, Sukadev Meher · 2022 8th International Conference on Signal Processing and Communication (ICSC) · 2022
A pre-processing interpolation method is suggested for producing edge preserved high resolution (HR) images of greater size from low resolution (LR) images. There are several interpolation techniques that cause non-uniform blurring in the upscaled image. Upscaling, like low pass filtering, produces blurring in the image’s edge region when compared to the flat region. Prior to upscaling, a pre-processing scheme is applied to mitigate blurring artifacts caused by interpolation in order to recover fine details and high frequency (HF). Using a combination of low pass and high pass filters, the LR is decomposed into various frequency sub-bands. The degraded HF bands are extracted and sharpened using higher order laplacian filters (HOLF). This method restores the HF details prior to interpolation to produce an HR image with no blurring artifacts. According to objective and subjective measures, the proposed algorithm outperforms other state-of-the-art methods.