AbstractA dataset of 439 confocal laser scanning microscopic images was analyzed to investigate the potential of an image-based automated analysis for identifying and assessing adherent thrombocytes on polymer surfaces. Parameters for image optimization of glutardialdehyde induced fluorescence images were classified and data mining was performed using the Java image processing software ImageJ. Previously reported analysis required that each thrombocyte had to be identified interactively and outlined manually. Now, we were able to determine the number and area of adherent thrombocytes with high accuracy (spearman correlation coefficient r = 0.98 and r = 0.99) using a two-stage filter-set, including a rolling ball background subtraction- and a watershed segmentation-algorithm. Furthermore, we could proof a significant correlation between these parameters (spearman correlation coefficient r = 0.97), determining both as suitable predictors for the evaluation of material induced thrombogenicity. The here reported image-based automated analysis can be successfully applied to identify and measure adherent thrombocytes on polymer surfaces and, thus, might be successfully integrated in a high-throughput screening process to evaluate biomaterial hemocompatibility.