Why don’t you see what I mean? Prospects and limitations of current automatic sign recognition research

This article presents an overview of current automatic sign recognition research. Based on a review of recent studies as well as on own research, several problem areas are identified that hamper the successful recognition of signed utterances by a computer. Some of these problems are shared with automatic speech recognition, while others seem to be particular to automatic sign recognition. These latter problems include contextdependency, determining the basic units of modeling, distinguishing signs from gestures, movement epenthesis and repetition within signs. As a possible solution to these problems, it is suggested that bottom-up processing should be supplemented with topdown processing.