Automated hand gesture recognition exploiting Active Learning methods

Automated hand gesture recognition is a trend that comes from the mixture of Machine Learning, Human Computer Interaction and Computer Vision fields. Its integration with several aspects of daily life has been proven exceptionally effective, improving the corresponding applications and their availability to larger groups of people. Although such systems are constructed based on generic rules, extra user-based information could be provided for tuning reasons. This work exploits Active Learning theory for enhancing the learning ability of the implemented hand gesture recognition system examining several supervised algorithms. The proposed framework does not demand neither much computational resources nor special equipment, favoring its adoption by both academic institutes and domestic users.1