Title: "Wavelet Filtering for Automated Searches of Gravitational Arcs"

Abstract:

Because of the dimness of the arcs and the poor signal to noise of the SDSS images, it is very important to filter out as much noise as possible before performing an arc search. Because the arcs are very dim, simple cuts on the image brightness will most likely eliminate the arc from the image. Also, because the arc images will most likely be located near bright galaxy clusters, it is important to not eliminate the noise by averaging over neighboring pixels because the arc image will then merge into the bright galaxy pixels making it difficult to detect. Because the arc is an extended object, noise pixels can be removed by using wavelet techniques. Wavelet techniques are superior to Fourier techniques because the noise is uncorrelated over large scales and the arc extends over many pixels.