Cell Tracking Challenge (www.celltrackingchallenge.net) is an internationally renowned competition aimed at objective benchmarking of state-of-the-art cell tracking algorithms. The preparation of reference tracking annotations used for performance evaluation is currently a laborious process for the organizers of this competition because it requires manual detection and correction of inconsistencies in triplets of manual annotations created by human experts. The primary objective of this thesis is to become familiar with the annotation protocol and tool used in the Cell Tracking Challenge and develop an image-analysis-driven routine for automatic detection of inconsistencies between the triplets of manual tracking annotations and for their semi-automatic correction, thus facilitating the preparation of reference tracking annotations. The developed routine will be integrated into the existing annotation tool written in C++ and quantitatively evaluated over the real datasets of different microscopy modalities included in the Cell Tracking Challenge.