algorithm is applied between two and four billion times a day, mainly on digital
pictures taken by smartphones and digital cameras. The main design driver for
the algorithm was to assure a compression without a significant loss of the
visual quality of the pictures. This is,
of course, basically a subjective requirement.
The purpose of this project is to evaluate the performance of the JPEG
algorithm following more objective criteria.
for this investigation will be a set of deep broad band images, obtained with
the Hubble Space Telescope. These images have been analysed in great detail
exploiting the SExtractor software package. The performance of JPEG will be
obtained by comparing the SExtractor results before and after the compression
of the images by JPEG. The basic quantities to compare will be the point spread
functions, the photometric accuracies and detection limits. The feasibility for
applying JPEG on astronomical digital images will be discussed.
I samarbejde medBirger Niss, NNIT A/S