Summarization of human activity videos via low-rank approximation

Ioannis Mademlis; Anastasios Tefas; Nikos Nikolaidis; Ioannis Pitas

doi:10.1109/ICASSP.2017.7952432

Summarization of human activity videos via low-rank approximation

Ioannis Mademlis, Anastasios Tefas, Nikos Nikolaidis, Ioannis Pitas

Department of Electrical & Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

11 Citations (Scopus)

431 Downloads (Pure)

Abstract

Summarization of videos depicting human activities is a timely problem with important applications, e.g., in the domains of surveillance or film/TV production, that steadily becomes more relevant. Research on video summarization has mainly relied on global clustering or local (frame-by-frame) saliency methods to provide automated algorithmic solutions for key-frame extraction. This work presents a method based on selecting as key-frames video frames able to optimally reconstruct the entire video. The novelty lies in modelling the reconstruction algebraically as a Column Subset Selection Problem (CSSP), resulting in extracting key-frames that correspond to elementary visual building blocks. The problem is formulated under an optimization framework
and approximately solved via a genetic algorithm. The proposed video summarization method is being evaluated using a publicly available
annotated dataset and an objective evaluation metric. According to the
quantitative results, it clearly outperforms the typical clustering approach.

Original language	English
Title of host publication	2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017)
Subtitle of host publication	Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	1627-1631
Number of pages	5
ISBN (Electronic)	9781509041176
ISBN (Print)	9781509041183
DOIs	https://doi.org/10.1109/ICASSP.2017.7952432
Publication status	Published - Aug 2017

Publication series

Name
ISSN (Print)	2379-190X

Keywords

video summarization
Sparse dictionary learning
Genetic algorithm

Access to Document

10.1109/ICASSP.2017.7952432

Full-text PDF (accepted author manuscript)
This is the author accepted manuscript (AAM). The final published version (version of record) is available online via IEEE at http://ieeexplore.ieee.org/document/7952432/. Please refer to any applicable terms of use of the publisher.
Accepted author manuscript, 2.51 MB

Handle.net

Persistent link

Cite this

Mademlis, I., Tefas, A., Nikolaidis, N., & Pitas, I. (2017). Summarization of human activity videos via low-rank approximation. In 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017): Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA (pp. 1627-1631). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/ICASSP.2017.7952432

Mademlis, Ioannis ; Tefas, Anastasios ; Nikolaidis, Nikos et al. / Summarization of human activity videos via low-rank approximation. 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017): Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA. Institute of Electrical and Electronics Engineers (IEEE), 2017. pp. 1627-1631

@inproceedings{ef20651720ee4beb8dce5bc98f70871f,

title = "Summarization of human activity videos via low-rank approximation",

abstract = "Summarization of videos depicting human activities is a timely problem with important applications, e.g., in the domains of surveillance or film/TV production, that steadily becomes more relevant. Research on video summarization has mainly relied on global clustering or local (frame-by-frame) saliency methods to provide automated algorithmic solutions for key-frame extraction. This work presents a method based on selecting as key-frames video frames able to optimally reconstruct the entire video. The novelty lies in modelling the reconstruction algebraically as a Column Subset Selection Problem (CSSP), resulting in extracting key-frames that correspond to elementary visual building blocks. The problem is formulated under an optimization frameworkand approximately solved via a genetic algorithm. The proposed video summarization method is being evaluated using a publicly availableannotated dataset and an objective evaluation metric. According to thequantitative results, it clearly outperforms the typical clustering approach.",

keywords = "video summarization, Sparse dictionary learning, Genetic algorithm",

author = "Ioannis Mademlis and Anastasios Tefas and Nikos Nikolaidis and Ioannis Pitas",

year = "2017",

month = aug,

doi = "10.1109/ICASSP.2017.7952432",

language = "English",

isbn = "9781509041183",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

pages = "1627--1631",

booktitle = "2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017)",

address = "United States",

}

Mademlis, I, Tefas, A, Nikolaidis, N & Pitas, I 2017, Summarization of human activity videos via low-rank approximation. in 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017): Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA. Institute of Electrical and Electronics Engineers (IEEE), pp. 1627-1631. https://doi.org/10.1109/ICASSP.2017.7952432

Summarization of human activity videos via low-rank approximation. / Mademlis, Ioannis; Tefas, Anastasios; Nikolaidis, Nikos et al.
2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017): Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA. Institute of Electrical and Electronics Engineers (IEEE), 2017. p. 1627-1631.

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

TY - GEN

T1 - Summarization of human activity videos via low-rank approximation

AU - Mademlis, Ioannis

AU - Tefas, Anastasios

AU - Nikolaidis, Nikos

AU - Pitas, Ioannis

PY - 2017/8

Y1 - 2017/8

N2 - Summarization of videos depicting human activities is a timely problem with important applications, e.g., in the domains of surveillance or film/TV production, that steadily becomes more relevant. Research on video summarization has mainly relied on global clustering or local (frame-by-frame) saliency methods to provide automated algorithmic solutions for key-frame extraction. This work presents a method based on selecting as key-frames video frames able to optimally reconstruct the entire video. The novelty lies in modelling the reconstruction algebraically as a Column Subset Selection Problem (CSSP), resulting in extracting key-frames that correspond to elementary visual building blocks. The problem is formulated under an optimization frameworkand approximately solved via a genetic algorithm. The proposed video summarization method is being evaluated using a publicly availableannotated dataset and an objective evaluation metric. According to thequantitative results, it clearly outperforms the typical clustering approach.

AB - Summarization of videos depicting human activities is a timely problem with important applications, e.g., in the domains of surveillance or film/TV production, that steadily becomes more relevant. Research on video summarization has mainly relied on global clustering or local (frame-by-frame) saliency methods to provide automated algorithmic solutions for key-frame extraction. This work presents a method based on selecting as key-frames video frames able to optimally reconstruct the entire video. The novelty lies in modelling the reconstruction algebraically as a Column Subset Selection Problem (CSSP), resulting in extracting key-frames that correspond to elementary visual building blocks. The problem is formulated under an optimization frameworkand approximately solved via a genetic algorithm. The proposed video summarization method is being evaluated using a publicly availableannotated dataset and an objective evaluation metric. According to thequantitative results, it clearly outperforms the typical clustering approach.

KW - video summarization

KW - Sparse dictionary learning

KW - Genetic algorithm

U2 - 10.1109/ICASSP.2017.7952432

DO - 10.1109/ICASSP.2017.7952432

M3 - Conference Contribution (Conference Proceeding)

SN - 9781509041183

SP - 1627

EP - 1631

BT - 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017)

PB - Institute of Electrical and Electronics Engineers (IEEE)

ER -

Mademlis I, Tefas A, Nikolaidis N, Pitas I. Summarization of human activity videos via low-rank approximation. In 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2017): Proceedings of a meeting held 5-9 March 2017, New Orleans, Louisiana, USA. Institute of Electrical and Electronics Engineers (IEEE). 2017. p. 1627-1631 Epub 2017 Jun 19. doi: 10.1109/ICASSP.2017.7952432

Summarization of human activity videos via low-rank approximation

Abstract

Publication series

Keywords

Access to Document

Handle.net

Fingerprint

Cite this