Home // International Journal On Advances in Systems and Measurements, volume 8, numbers 3 and 4, 2015 // View article

Accuracy Evaluation of Second-Order Shape Prediction on Tracking Non-Rigid Objects

Authors:
Kenji Nishida
Takumi Kobayashi
Jun Fujiki

Keywords: Tracking non-rigid objects; Chamfer distance; Shape prediction; Optical flow.

Abstract:
For our previously proposed shape prediction based tracking algorithm for non-rigid objects, the shape prediction accuracy is critical for the tracking performance. Therefore, we have presented a preliminary evaluation of second-order shape prediction algorithm for tracking non-rigid objects. In the proposed algorithm, the object shape was predicted from the movement of feature points, which were approximated by a second-order Taylor expansion. Approximate first-order movements, the so-called optical flows, were simultaneously exploited by chamfer matching of edgelets. Shape prediction accuracy was evaluated by chamfer matching between the predicted object shape and the actual object shape. While only one video sequence was preliminary evaluated, three more video sequences were evaluated. The new sequences are captured by fixed camera, while our previous sequence was captured by Hand-held camera. In this paper, the effect of second-order shape prediction is quantitatively analyzed by more video sequences. The method exhibits superior shape prediction performance compared to a simple linear prediction method.

Pages: 201 to 209

Copyright: Copyright (c) to authors, 2015. Used with permission.

Publication date: December 30, 2015

Published in: journal

ISSN: 1942-261x