Usability improvement of skeletal-based gesture recognition for exploring virtual environments


Babaei, Mahdi (2015) Usability improvement of skeletal-based gesture recognition for exploring virtual environments. Masters thesis, Multimedia University.

Full text not available from this repository.


The aim of this thesis is to improve the usability of gesture recognition based on a skeleton tracking method for virtual environments. Gesture recognition has been considered to be as one of the most effective input methods (based on the human’s natural body movements) to interact with virtual environments naturally (Rautaray and Agrawal, 2011). The skeleton tracking techniques that have been widely used so far for this purpose have shown common accuracy issues in recognising Micro-gestures. To resolve the disadvantages associated with skeleton tracking, the author designed a theoretical framework that applies Vygotski’s (1987) Activity theory on the taxonomy of usability characteristics of virtual environments by Gabbard (1997) to fill in the gap of Norman’s (1988) Theory of Action. The outcome will study the tasks in virtual environments in three levels according to the activity theory that makes it easier to troubleshoot the accuracy problems. A technical framework has been designed that benefits from a wearable multimodal gesture recognition input device that can work simultaneously with skeleton tracking devices. A two-iteration usability test has been performed to enhance the user-centred design of a gesture-based interface using the proposed combination of modalities.

Item Type: Thesis (Masters)
Additional Information: Call No.: TK7882.P3 M34 2015
Uncontrolled Keywords: Pattern recognition systems (Electronics)
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800-8360 Electronics
Divisions: Faculty of Creative Multimedia (FCM)
Depositing User: Ms Nurul Iqtiani Ahmad
Date Deposited: 19 Jan 2016 08:13
Last Modified: 19 Jan 2016 08:13


Downloads per month over past year

View ItemEdit (login required)