Browsing by Author "Achmed, Imran"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Improved Hand-Tracking Framework with a Recovery Mechanism
    (Telkom, 2013) Achmed, Imran; Venter, Isabella M.; Eisert, Peter
    Abstract−Hand-tracking is fundamental to translating sign language to a spoken language. Accurate and reliable sign language translation depends on effective and accurate hand-tracking. This paper proposes an improved hand-tracking framework that includes a tracking recovery algorithm optimising a previous framework to better handle occlusion. It integrates the tracking recovery algorithm to improve the discrimination between hands and the tracking of hands. The framework was evaluated on 30 South African Sign Language phrases that use: a single hand; both hands without occlusion; and both hands with occlusion. Ten individuals in constrained and unconstrained environments performed the gestures. Overall, the proposed framework achieved an average success rate of 91.8% compared to an average success rate of 81.1% using the previous framework. The results show an improved tracking accuracy across all signs in constrained and unconstrained environments.
  • Thumbnail Image
    Item
    Upper body pose recognition and estimation towards the translation of South African sign language
    (University of the Western Cape, 2011) Achmed, Imran; Connan, James; Dept. of Computer Science; Faculty of Economics and Management Sciences
    Recognising and estimating gestures is a fundamental aspect towards translating from a sign language to a spoken language. It is a challenging problem and at the same time, a growing phenomenon in Computer Vision. This thesis presents two approaches, an example-based and a learning-based approach, for performing integrated detection, segmentation and 3D estimation of the human upper body from a single camera view. It investigates whether an upper body pose can be estimated from a database of exemplars with labelled poses. It also investigates whether an upper body pose can be estimated using skin feature extraction, Support Vector Machines (SVM) and a 3D human body model. The example-based and learning-based approaches obtained success rates of 64% and 88%, respectively. An analysis of the two approaches have shown that, although the learning-based system generally performs better than the example-based system, both approaches are suitable to recognise and estimate upper body poses in a South African sign language recognition and translation system.