Fig. 9
From: A richly annotated dataset of co-speech hand gestures across diverse speaker contexts
Quality of lexeme groups based on similarity of annotations’ physical features - BOTH. (a). Pair-wise (dis-)similarity of lexemes (mean DTW distances across annotations). Darker colors indicate higher similarity. (b). Rank-similarity based on DTW values (mean across annotations), ordered from left (highest similarity) to right (lowest similarity). Higher scores correspond to higher within-group similarity.
