Abstract:A number of problems in computer vision and related fields would be mitigated if camera spectral sensitivities were known. As consumer cameras are not designed for high-precision visual tasks, manufacturers do not disclose spectral sensitivities. Their estimation requires a costly optical setup, which triggered researchers to come up with numerous indirect methods that aim to lower cost and complexity by using color targets. However, the use of color targets gives rise to new complications that make the estimation more difficult, and consequently, there currently exists no simple, low-cost, robust go-to method for spectral sensitivity estimation. Furthermore, even if not limited by hardware or cost, researchers frequently work with imagery from multiple cameras that they do not have in their possession. To provide a practical solution to this problem, we propose a framework for spectral sensitivity estimation that not only does not require any hardware, but also does not require physical access to the camera itself. Similar to other work, we formulate an optimization problem that minimizes a two-term objective function: a camera-specific term from a system of equations, and a universal term that bounds the solution space. Different than other work, we use publicly available high-quality calibration data to construct both terms. We use the colorimetric mapping matrices provided by the Adobe DNG Converter to formulate the camera-specific system of equations, and constrain the solutions using an autoencoder trained on a database of ground-truth curves. On average, we achieve reconstruction errors as low as those that can arise due to manufacturing imperfections between two copies of the same camera. We provide our code and predicted sensitivities for 1,000+ cameras, and discuss which tasks can become trivial when camera responses are available.

Energy balance in Spectral Filter Array camera design

Fabry-Perot Multilayer for Spectral Imaging Applications

Design of Compact Wide-Band Triple Grating Imaging Spectrometer Based on Freeform Surfaces

Color Router-based Long-wave Infrared Multispectral Imaging

Design of a Dual-Mode Multispectral Filter Array

Fast Multispectral Imaging by Spatial Pixel-Binning and Spectral Unmixing

A Single Sensor Based Multispectral Imaging Camera using a Narrow Spectral Band Colour Mosaic Integrated on the Monochrome CMOS Image Sensor

Designing Color Filters that Make Cameras MoreColorimetric

Practical spectral photography II: snapshot spectral imaging using linear retarders and microgrid polarization cameras

A single sensor based multispectral imaging camera using a narrow spectral band color mosaic integrated on the monochrome CMOS image sensor

Spectral DiffuserCam: lensless snapshot hyperspectral imaging with a spectral filter array

Ultrabroadband, high color purity multispectral color filter arrays

Spectral Sensitivity Estimation Without a Camera

Hyperpixels: Pixel Filter Arrays of Multivariate Optical Elements for Optimized Spectral Imaging

Grayscale-to-color: Single-step fabrication of bespoke multispectral filter arrays

Opti-MSFA: a toolbox for generalized design and optimization of multispectral filter arrays

A light-efficient and versatile multiplexing method for snapshot spectral imaging

Multispectral color image capture using a liquid crystal tunable filter

Multispectral imaging via nanostructured random broadband filtering

Automatic Design of High-Sensitivity Color Filter Arrays with Panchromatic Pixels