Multi-Channel Near-Infrared Bandpass Mosaic Filter for Spectral Fundus Imaging

Honghao Tang, Hironari Takehara, Yusuke Horiki, Makito Haruta, Kiyotaka Sasagawa, Jun Ohta, Hiroyuki Tashiro

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We previously reported a near-infrared (NIR) color fundus camera that can image the fundus without the use of mydriatic agents. Using this technology, pseudo-color images can be reconstructed from images in the three NIR wavelength bands. In this study, we designed and manufactured a five-wavelength bandpass mosaic filter for CMOS image sensors (CIS) by modifying the thickness of the middle layer of a multilayer thin-film structure based on the interference bandpass filter principle. It is possible to acquire information about all bands at all locations on the fundus by using the fixational movement of the human eye. Adapting a fabricated 5-channel CIS, we report the validity of high-spatial-resolution imaging by demosaicing and spectral resolution enhancement by deconvolution processing.

Original languageEnglish
Title of host publication2022 IEEE International Meeting for Future of Electron Devices, Kansai, IMFEDK 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665474535
DOIs
Publication statusPublished - 2022
Event2022 IEEE International Meeting for Future of Electron Devices, Kansai, IMFEDK 2022 - Kyoto, Japan
Duration: Nov 28 2022Nov 30 2022

Publication series

Name2022 IEEE International Meeting for Future of Electron Devices, Kansai, IMFEDK 2022

Conference

Conference2022 IEEE International Meeting for Future of Electron Devices, Kansai, IMFEDK 2022
Country/TerritoryJapan
CityKyoto
Period11/28/2211/30/22

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering
  • Electronic, Optical and Magnetic Materials
  • Instrumentation

Fingerprint

Dive into the research topics of 'Multi-Channel Near-Infrared Bandpass Mosaic Filter for Spectral Fundus Imaging'. Together they form a unique fingerprint.

Cite this