# Code and Data for "Interpolant-based demosaicing routines for dual-mode visible/near-infrared imaging systems" **Paper Title:** Interpolant-based demosaicing routines for dual-mode visible/near-infrared imaging systems **Paper Authors:** Steven Blair and Viktor Gruev **Paper Abstract**: Dual-mode visible/near-infrared imaging systems, including a bioinspired six-channel design and more conventional four-channel implementations, have transitioned from a niche in surveillance to general use in machine vision. However, the demosaicing routines that transform the raw images from these sensors into processed images that can be consumed by humans or computers rely on assumptions that may not be appropriate when the two portions of the spectrum contribute different information about a scene. A solution can be found in a family of demosaicing routines that utilize interpolating polynomials and splines of different dimensionalities and orders to process images with minimal assumptions. **Journal:** Optics Express **Year:** 2022 ## The Structure of the Dataset All code has been uploaded to Zenodo, and all data has been uploaded to Dryad. The two repositories should be linked to each other on both the Zenodo website and the Dryad website. Details related to the data are described in this README; details related to the code are described in a different README provided on the Zenodo website. The data for the single-mode dataset consists of a Tag Image File Format (TIF) file (with the extension ".tif") which can be manually opened with, e.g., the image viewer "IrfanView" (as provided by Irfan Å kiljan) or programmatically opened with, e.g., the Python library "Pillow" (as developed by Fredrik Lundh and other contributors). The TIF file contains a full-color image, along with the associated metadata, that offers up a visible depiction of a scene. The data for the dual-mode dataset consists of Hierarchical Data Format (HDF) files (with the extension ".h5") which can be manually opened with, e.g., the software "HDFView" (as provided by the The HDF Group) or programmatically opened with, e.g., the Python library "h5py" (as developed by Andrew Collette and other contributors). Each H5 file contains a three-channel image, along with the associated metadata; the H5 files have been divided up, first, according to the primary subject of the image ("Image 01" - "Image 10") and, second, according to whether the image contains a visible depiction of a scene ("VIS - Scene"), a visible depiction of a color chart ("VIS - Color Chart"), a near-infrared depiction of a scene ("NIR - Scene"), or a near-infrared depiction of a color chart ("NIR - Color Chart").