Nowadays, pedestrian tracking is an important task in such fields as autonomous driving, sport analytics and video surveilence. Although different Machine Learning and Deep Learning approaches thrived in this field they still need a lot of labeled data to be trained on. Manual data labeling becomes a new bottleneck in scaling up supervised learning approaches. In this simple mark-up tool we try to speed up performance of manual data labeling using semi-automation based on CNN networks and algorithmic track processing, but still leaving an opportunity to manually refine the results of the automated tracking.
