Automatic recognition of hand shapes in a moving image sequence requires the elements of handtracking, feature extraction and classification. We have developed a robust tracking algorithm and a newhand shape representation technique that characterises the finger-only topology of the hand by adaptingan existing technique from speech signal processing. The tracking algorithm determines the centre ofthe largest convex subset of the hand throughout an image sequence, using a combination of patternmatching and condensation algorithms. A hand shape feature represents the topological formationof the finger-only regions of the hand using a Linear Predictive Coding parameter set called cepstralcoefficients. Feature extraction is performed on the polar dimensions of the hand region-of-interest, bytracking the finger-only region and extracting Euclidean distances between the finger-only contour andthe hand centre, which are then converted into cepstral coefficients. Experiments are conducted usingmug-grabbing sequences to recognise 4 different hand shapes. Results demonstrate the robustness ofhand tracking on cluttered backgrounds and the effectiveness of the hand shape representation technique on varying hand shapes.
|Journal||Machine Graphics and Vision|
|Publication status||Published - 2003|