WHAT DO YOU GAIN FROM DECONVOLUTION - OBSERVING FAINT GALAXIES WITH THE HUBBLE-SPACE-TELESCOPE WIDE FIELD CAMERA

We describe experiments with deconvolutions of simulations of deep Hubble Space Telescope Wide Field Camera images containing faint, compact galaxies. Our aim is to determine under what circumstances there is a quantitative advantage to image deconvolution. In particular, we explore whether it is helpful for distinguishing between stars and compact galaxies, or between spiral and elliptical galaxies, and whether it improves the accuracy with which characteristic radii and integrated magnitudes may be determined. We use both the Maximum Entropy and Richardson-Lucy deconvolution algorithms, and find that they give the same results. For medium and low signal-to-noise images, deconvolution does not significantly improve our ability to distinguish between faint stars and compact galaxies, nor between spiral and elliptical galaxies. Quantitative measurements from both raw and deconvolved images are biased, and corrections to these biases must be determined. It is easier to quantify and remove the biases from measurements made from images that have not been deconvolved. We find no benefit from deconvolution for measuring luminosity profiles but we emphasize that these results are limited to low signal-to-noise ratio images of very compact (often undersampled) galaxies.