Abstract
This thesis work aimed to reimplement and improve an existing system for hand pose recognition from monocular video data. The resulting system is light, multi-platform and easily extensible because of its modularity. It relies on treating the problem of hand pose estimation as a nearest neighbour look-up in a database of synthetically generated hand images. Its main characteristics are the use of HOGs (Histogram of Oriented Gradients) as features and employing temporal consistency for greater reliability and robustness. The paper also makes a review of the current hand pose recognition research and gives arguments for our choices of implementation both in terms of design and actual technology used.