High-Resolution Metalens Imaging with Sequential Artificial Intelligence Models

Wei Lun Hsu, Chen Fu Huang, Chih Chun Tan, Noreena Yi Chin Liu, Cheng Hung Chu, Po Sheng Huang, Pin Chieh Wu, Shang Jyh Yiin, Takuo Tanaka, Chun Jen Weng, Chih Ming Wang

研究成果: 雜誌貢獻期刊論文同行評審

1 引文 斯高帕斯(Scopus)

摘要

An analysis of the optical response of a GaN-based metalens was conducted alongside the utilization of two sequential artificial intelligence (AI) models in addressing the occasional issues of blurriness and color cast in captured images. The optical loss of the metalens in the blue spectral range was found to have resulted in the color cast of images. Autoencoder and CodeFormer sequential models were employed in order to correct the color cast and reconstruct image details, respectively. Said sequential models successfully addressed the color cast and reconstructed details for all of the allocated face image categories. Subsequently, the CIE 1931 chromaticity diagrams and peak signal-to-noise ratio analysis provided numerical evidence of the AI models’ effectiveness in image reconstruction. Furthermore, the AI models can still repair the image without blue information. Overall, the integration of metalens and artificial intelligence models marks a breakthrough in enhancing the performance of full-color metalens-based imaging systems.

原文???core.languages.en_GB???
頁(從 - 到)11614-11620
頁數7
期刊Nano Letters
23
發行號24
DOIs
出版狀態已出版 - 27 12月 2023

指紋

深入研究「High-Resolution Metalens Imaging with Sequential Artificial Intelligence Models」主題。共同形成了獨特的指紋。

引用此