Presentation Information
[17p-PA1-13]Luminance-Only Restormer Restoration for Noise Suppression in LC Lens Images
〇Ryoya Takewaki1, Sota Shimizu2, Marenori Kawamura1 (1.Akita Univ., 2.Aichi Univ. of Tech.)
Keywords:
liquid crystal lens,liquid crystal,image processing
Images captured with liquid crystal lenses may suffer from image degradation such as blur and noise due to optical aberrations and non-ideal phase distributions, resulting in reduced visual quality. In this study, a simple post-processing method is investigated to improve the quality of such images. The proposed approach applies a Restormer-based defocus deblurring process only to the luminance component obtained from the YCrCb color space, while noise suppression is applied to the chrominance components. By separating luminance restoration and color noise reduction, the method aims to suppress noise while preserving structural features without introducing color artifacts. Experimental results using images captured with a liquid crystal lens show a reduction in the no-reference image quality metric NIQE and an increase in edge strength, indicating effective noise suppression and preservation of image structures. These results suggest that the proposed luminance-focused restoration approach is a practical post-processing technique for improving image quality.