Rating a compression algorithm’s performance is usually done in experimental studies, where researchers have frequently used JPEG pre-compressed data. It is not clear yet, whether results of such compression experiments are reliable if conducted from pre-compressed data.
To investigate this issue, we study the impact of using pre-compressed data on iris segmentation and evaluate the relation between iris segmentation performance and general quality metrics. Furthermore we propose a method to overcome potential problems in case using pre-compressed data sets cannot be avoided, e.g. for reasons of ground-truth availability.