The color demosaicing and image scaling based on improve Hamilton-Adams

• Published in: Proceedings ... Asia-Pacific Signal and Information Processing Association Annual Summit and Conference APSIPA ASC ... (Online) pp. 892 - 897
• Main Authors: Peng, Yu-Wen, Hu, Chia-Yu, Chin, Yen-Ju, Chou, He-Sheng, Lin, Yuan-Jin, Liu, Yu-Lin, Chen, Shih-Lun, Chen, Tsung-Yi, Li, Kuo-Chen, Chen, Chiung-An
• Format: Conference Proceeding
• Language: English
• Published: IEEE 31.10.2023
• Subjects: Color filter array; Demosaicing; Hamilton-Adams; Image color analysis; Image edge detection; Image pre-processing; Image quality; Image scaling; Interpolation; Measurement; Noise reduction; Pipelines
• ISSN: 2640-0103
• DOI: 10.1109/APSIPAASC58517.2023.10317437

The demosaicing process is the most critical step in image processing, this study presents a Hamilton-Adams-based algorithm for mosaic removal and image scaling. The algorithm consists of three steps: image pre-processing, image scaling, and improved Hamilton-Adams mosaic removal. The image pre-processing stage employs a mean filter to reduce high-frequency signals at the image edges, effectively mitigating the zipper effect caused by image scaling and improving image quality. The improved Hamilton-Adams method incorporates high-quality linear interpolation to achieve more accurate edge computations. In terms of image quality assessment, this paper increased 1.13 dB of PSNR, 1.19 dB of CPSNR, and 0.03 of SSIM values compared to high-quality linear interpolation with bicubic interpolation.