Instance features based deep learning methods prompt the performances of high speed object tracking systems by directly comparing target with its template during training and tracking. However, from the perspective of human vision system, prior knowledge of target also plays key role during the process of tracking. To integrate both semantic knowledge and instance features, we propose a convolutional network based object tracking framework to simultaneously output bounding boxes based on different prior knowledge as well as confidences of corresponding Assumptions. Experimental results show that our proposed approach retains both higher accuracy and efficiency than other leading methods on tracking tasks covering most daily objects.
Suofei ZHANG
Nanjing University of Posts and Telecommunications
Bin KANG
Nanjing University of Posts and Telecommunications
Lin ZHOU
Southeast University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Suofei ZHANG, Bin KANG, Lin ZHOU, "Object Tracking by Unified Semantic Knowledge and Instance Features" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 3, pp. 680-683, March 2019, doi: 10.1587/transinf.2018EDL8181.
Abstract: Instance features based deep learning methods prompt the performances of high speed object tracking systems by directly comparing target with its template during training and tracking. However, from the perspective of human vision system, prior knowledge of target also plays key role during the process of tracking. To integrate both semantic knowledge and instance features, we propose a convolutional network based object tracking framework to simultaneously output bounding boxes based on different prior knowledge as well as confidences of corresponding Assumptions. Experimental results show that our proposed approach retains both higher accuracy and efficiency than other leading methods on tracking tasks covering most daily objects.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2018EDL8181/_p
Copy
@ARTICLE{e102-d_3_680,
author={Suofei ZHANG, Bin KANG, Lin ZHOU, },
journal={IEICE TRANSACTIONS on Information},
title={Object Tracking by Unified Semantic Knowledge and Instance Features},
year={2019},
volume={E102-D},
number={3},
pages={680-683},
abstract={Instance features based deep learning methods prompt the performances of high speed object tracking systems by directly comparing target with its template during training and tracking. However, from the perspective of human vision system, prior knowledge of target also plays key role during the process of tracking. To integrate both semantic knowledge and instance features, we propose a convolutional network based object tracking framework to simultaneously output bounding boxes based on different prior knowledge as well as confidences of corresponding Assumptions. Experimental results show that our proposed approach retains both higher accuracy and efficiency than other leading methods on tracking tasks covering most daily objects.},
keywords={},
doi={10.1587/transinf.2018EDL8181},
ISSN={1745-1361},
month={March},}
Copy
TY - JOUR
TI - Object Tracking by Unified Semantic Knowledge and Instance Features
T2 - IEICE TRANSACTIONS on Information
SP - 680
EP - 683
AU - Suofei ZHANG
AU - Bin KANG
AU - Lin ZHOU
PY - 2019
DO - 10.1587/transinf.2018EDL8181
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 3
JA - IEICE TRANSACTIONS on Information
Y1 - March 2019
AB - Instance features based deep learning methods prompt the performances of high speed object tracking systems by directly comparing target with its template during training and tracking. However, from the perspective of human vision system, prior knowledge of target also plays key role during the process of tracking. To integrate both semantic knowledge and instance features, we propose a convolutional network based object tracking framework to simultaneously output bounding boxes based on different prior knowledge as well as confidences of corresponding Assumptions. Experimental results show that our proposed approach retains both higher accuracy and efficiency than other leading methods on tracking tasks covering most daily objects.
ER -