Go to | Data for computer vision | Kobus Barnard 's research page | Kobus Barnard 's Home Page


Fluorescent Surface Reflectance Spectra



The official source for this data is elsewhere, and the appropriate web resource reference to the data is one of
http://www.cs.sfu.ca/~colour/data/index.html
http://www.cs.sfu.ca/~colour/data/colour_constancy_synthetic_test_data/index.html
However, the associated meta data is likely to be corrected/updated more frequently in this version of the interface.

Questions, comments, and problems with this data should be directed to Kobus Barnard


This directory contains some the data presented in:

Kobus Barnard, Lindsay Martin, Brian Funt, and Adam Coath, " Data for Colour Research," Color Research and Application, Volume 27, Issue 3, pp. 148-152, 2002.

(The appropriate archival reference for this data).


This data in this directory is used in the following publications:

Kobus Barnard, " Color Constancy with Fluorescent and Surfaces," in preparation.

Chapter 7 of "Practical Colour Constancy," Phd thesis, Simon Fraser University, School of Computing (1999)

Kobus Barnard, "Color Constancy with Fluorescent and Surfaces," Proceedings of the IS&T/SID Seventh Color Imaging Conference: Color Science, Systems and Applications, 1999, pp. 257-261.


The data was collected by Kobus Barnard, Lindsay Martin, and Adam Coath, under the guidance of Kobus Barnard in Brian Funt's Computational Colour Vision Laboratory.


Data Description

To describe the reflectance from fluorescent surfaces one normally uses a matrix which specifies the complete spectra emitted as the result of being exposed to a given wavelength. The matrix is triangular, as light never gets reflected at higher energy (smaller wavelength). Such data is hard to get a hold off, and difficult to measure, so we took a different approach. Since the dimensionality of comon illuminants is limited, we claim that one can characterize fluorescent surface reflectance adequately by simply specifying the response to a number of illuminants. Since we have a collection of sources and filters in our lab, this approach is much easier to execute.

The format of the data is simply the native format we use. We use the "#" character in the first column to denote a comment line. Lines beginning with "#!" have additional special meaning, but can be ignored, except the "#!eof" string, which is used as described below. Blanks lines also appear freely. Therefore, it is likely that some data massaging will be required to use this data.

The data file contains data for 11 candidate fluorescent spectra. The data for each surface is terminated with a "#!eof" in the first column. The data for each surface is a collection of pairs of spectra. Each spectra consists of 101 numbers, being the energy from 380nm to 780nm, in 4nm increments. Each pair is an input/output pair. The input is the spectral energy of the illumination, as measured by the reflected light from a white reflectance standard. The output spectra is the energy reflected back by the surface. If the surfaces were not fluorescent, then the information would be redundant, as the ratio of these two signals would always be the same (modulo noise and wavelengths of no energy). However, since they are fluorescent, the ratios are not all the same. Making these measurements with common illuminants allows us to characterize a fluorescent surface to some extent without needing to attempt to measure the complete triangular matrix of fluorescent reflection. This works because the dimensionality of common illuminants is limited. See the references listed above for further information.


The Data

Fluorescent reflectance data (gzipped ascii)