This item is available under a Creative Commons License for non-commercial use only
1.2 COMPUTER AND INFORMATION SCIENCE
A shape distribution is a histogram used to uniquely identify different shapes. The histogram is produced by taking random distances on the surface of a shape or object. Theoretically, each shape or object should produce a unique histogram, as the distribution of distances for each shape should be different. Shape distributions have recently been implemented in a number of object recognition areas. They are an attractive method as they are inherently simple, fast and generic. This paper presents the results of research undertaken on the application of shape distributions for the purpose of sign language recognition. There are four main elements that need to be undertaken in building a fully operational sign language recognition system, namely: posture, position, orientation and motion. It is the first of these components that this paper addresses.
Young, E. and Clynch. G. Hand posture recognition in sign language using shape distributions. 6th Annual Information Technology & Telecommunications (IT&T) Conference, Carlow, Ireland, 2006.