For this purpose, we have taken the relative response of all locations on the detector as 1. The current estimates are that the zero-points defined here are accurate to within +/- 10% (1 sigma).
Note that one GALEX count corresponds to one detected "average" photon for the (respective) bandpass. Since detector background is very small (less than 1%), GALEX counts may be used in Poisson statistics to compute S/N for sources or sky background. More details on GALEX counts and background may be found in the next section.
* A "dose" image, which is a picture of what the detector saw during the eclipse (i.e. little donuts of dithered source images). "Wiggle" and "walk" corrections have been applied to these images, and move counts around but do not scale them. The "dose" image has 3 arcsec pixels
* A "cnt" image, in which the stars looks like stars but which still has no scaling...so this map is still in detector count units. The "cnt" image has 1.5 arcsec pixels.
* An "int" image, which has had the relative response correction folded in: thus, the summed signal for a given source in this image is the proper quantity to use in determining a GALEX magnitude (i.e. take the log and scale). The "int" image has 1.5 arcsec pixels.
Typical GALEX background in the FUV is 2000 cps for the whole field, which corresponds to 3x10-4 cps/pixel (where a pipeline pixel is 1.5 arcsec). This is the typical total signal, and about 1/2 of it is detector background, so take the FUV detector background to be ~10-4 cps/pixel for simplicity. NUV detector background is 10x higher, or ~10-3 cps/pixel. if a typical FUV source (19 mag, or 1 cps) covers 9 pixels (5" FWHM), then the detector background is about 0.001 cps/source, compared to 1 cps of signal. Thus, the background would be ~0.1% of the typical souce for the the FUV. In the NUV, for a 20 mag source (1 cps), we're looking at a detector background of 10-3 cps/pixel, or 0.01 cps/source for a 5" FWHM, and in this case the NUV background is ~1% of the source flux.