Title:
OBJECT TRACKING METHOD, OBJECT TRACKING DEVICE, AND PROGRAM
Document Type and Number:
WIPO Patent Application WO/2018/030048
Kind Code:
A1
Abstract:
The present invention includes: an input step (S1) for inputting two or more images that are contiguous in time series to a neural network; and an output step (S2) for comparing the feature quantity of each of the two or more images inputted in the input step (S1), the one that is extracted to the neural network, to check similarity and thereby outputting, as an identification result, identification information and position information pertaining to the one or more objects that match one or more objects in an image preceding in time series, which are tracking candidates, and that are in an image following the preceding image in time series. The neural network includes two or more of the same structure having zero or more full bound layers and one or more convolution layers, and shares parameters in corresponding layers between the same structures.
Inventors:
KIM MIN YOUNG
TSUKIZAWA SOTARO
TSUKIZAWA SOTARO
Application Number:
PCT/JP2017/024932
Publication Date:
February 15, 2018
Filing Date:
July 07, 2017
Export Citation:
Assignee:
PANASONIC IP MAN CO LTD (JP)
International Classes:
G06T7/00; G06T7/246
Other References:
WANG X. ET AL.: "Unsupervised Learning of Visual Representations using Videos", IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV, 7 December 2015 (2015-12-07), pages 2794 - 2802, XP032866625
WANG B. ET AL.: "Joint Learning of Convolutional Neural Networks and Temporally Constrained Metrics for Tracklet Association", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW, 26 June 2016 (2016-06-26), pages 386 - 393, XP055557368
MITSURU ABE ET AL.: "Local feature description for keypoint matching", IEICE TECHNICAL REPORT, vol. 115, no. 388, 14 December 2015 (2015-12-14), pages 53 - 73, XP055557371
LAURA L. ET AL.: "Learning by tracking: Siamese CNN for robust target association", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW, 26 June 2016 (2016-06-26), pages 418 - 425, XP033027855
RAN T. ET AL.: "Siamese Instance Search for Tracking", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR, 27 June 2016 (2016-06-27), pages 1420 - 1429, XP033021318
GAN, Q.; GUO, Q.; ZHANG, Z.; CHO, K.: "First step toward model-free, anonymous object tracking with recurrent neural networks", CORR ABS/1511.06425
KAHOU, S.E.; MICHALSKI, V.; MEMISEVIC, R.: "RATM: recurrent attentive tracking model", CORR ABS/1510.08660, 2015
ONDRUSKA, P.; POSNER, I.: "Deep tracking: Seeing beyond seeing using recurrent neural networks", CORR ABS/1602.00991, 2016
WANG B. ET AL.: "Joint Learning of Convolutional Neural Networks and Temporally Constrained Metrics for Tracklet Association", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW, 26 June 2016 (2016-06-26), pages 386 - 393, XP055557368
MITSURU ABE ET AL.: "Local feature description for keypoint matching", IEICE TECHNICAL REPORT, vol. 115, no. 388, 14 December 2015 (2015-12-14), pages 53 - 73, XP055557371
LAURA L. ET AL.: "Learning by tracking: Siamese CNN for robust target association", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS (CVPRW, 26 June 2016 (2016-06-26), pages 418 - 425, XP033027855
RAN T. ET AL.: "Siamese Instance Search for Tracking", IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR, 27 June 2016 (2016-06-27), pages 1420 - 1429, XP033021318
GAN, Q.; GUO, Q.; ZHANG, Z.; CHO, K.: "First step toward model-free, anonymous object tracking with recurrent neural networks", CORR ABS/1511.06425
KAHOU, S.E.; MICHALSKI, V.; MEMISEVIC, R.: "RATM: recurrent attentive tracking model", CORR ABS/1510.08660, 2015
ONDRUSKA, P.; POSNER, I.: "Deep tracking: Seeing beyond seeing using recurrent neural networks", CORR ABS/1602.00991, 2016
Attorney, Agent or Firm:
KAMATA Kenji et al. (JP)
Download PDF:
Previous Patent: WIPER BLADE RUBBER
Next Patent: COMMUNICATION DEVICE, COMMUNICATION METHOD AND RECORDING MEDIUM
Next Patent: COMMUNICATION DEVICE, COMMUNICATION METHOD AND RECORDING MEDIUM