Textural feature descriptors for a static and dynamic hand gesture recognition system

Hand gesture recognition has become one of the most important directions in human-computer interaction (HCI) research. Despite recent advances in this area, the development of methods and techniques to correctly recognize gestures is still ongoing. In this paper, a hand gesture and sign recognition system (HGRS/SGRS) based on textural features is implemented using a local binary pattern (LBP), local directional pattern (LDP), local optimal-oriented pattern (LOOP) and local Gabor binary pattern histogram sequence (LGBPHS). In terms of feature extraction, we introduce Modified(i)-LOOP, a modified local texture descriptor for HGRS and SGRS to improve the efficiency of our system. The experiments are carried out on five datasets, for Arabic, American Alphabet Sign languages and dynamic gestures where the proposed M-i-LOOP as well as LGBPHS achieve satisfactory simulation results.