Optimal Subsequence Bijection

Abstract

We consider the problem of elastic matching of sequences of real numbers. Since

Correspondences between parts of shapes with particle filters

“…The data are shapes from the standard MPEG-7 similarity data set [9]. The data set provides 70 classes of 20 similar shapes each.…”

“…We compare our PF correspondence to part matching techniques based on time series algorithms described in [9]. As in [9], we took 10 parts of shapes of different classes from the MPEG-7 data set, the parts being identical to the ones described in [9].…”

“…As in [9], we took 10 parts of shapes of different classes from the MPEG-7 data set, the parts being identical to the ones described in [9]. We then matched each part to the entire database (all 1400 shapes), based on the PF correspondence (PFPA) and Procrustes Alignment.…”

“…We then matched each part to the entire database (all 1400 shapes), based on the PF correspondence (PFPA) and Procrustes Alignment. To achieve a fair comparison, we used curvature as shape descriptor, as in [9]. Fig.…”

“…how many of the top-n results are from the same class as the query part. PFPA is the described algorithm, OSB the algorithm described in [9]. DTWCW is a windowed version of Dynamic Time Warping, LCSS stands for Longest Common Sub Sequence.…”