DELFTGESTURES

Dissimilarity dataset.

    D = DELFTGESTURES

This dataset consists of the dissimilarities computed from a set of gestures  in a sign-language study. They are measured by two video cameras observing the  positions the two hands in 75 repititions of creating 20 different signs. The  dissimilarities result from a dynamic time warping procedure. The experiments  are performed by Gineke ten Holt, Jeroen Arendsen, Robbert Eggermont and Jeroen  Lichtenauer who also prepared the dataset.

Reference(s)

Jeroen Lichtenauer, Emile A. Hendriks, Marcel J. T. Reinders: Sign Language Recognition by Combining Statistical DTW and Independent Classification. IEEE Trans. Pattern Anal. Mach. Intell. vol. 30, 2040-2046, 2008.

See also

prtools, datasets, prdisdata,