Spatio-temporal mid-level feature bank for action recognition in low quality video

Saimunur Rahman; John See

doi:10.1109/ICASSP.2016.7471996

Spatio-temporal mid-level feature bank for action recognition in low quality video

Saimunur Rahman
, John See

School of Mathematical & Computer Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

7 Citations (Scopus)

Abstract

It is a great challenge to perform high level recognition tasks on videos that are poor in quality. In this paper, we propose a new spatio-temporal mid-level (STEM) feature bank for recognizing human actions in low quality videos. The feature bank comprises of a trio of local spatio-temporal features, i.e. shape, motion and textures, which respectively encode structural, dynamic and statistical information in video. These features are encoded into mid-level representations and aggregated to construct STEM. Based on the recent binarized statistical image feature (BSIF), we also design a new spatiotemporal textural feature that extracts discriminately from 3D salient patches. Extensive experiments on the poor quality versions/subsets of the KTH and HMDB51 datasets demonstrate the effectiveness of the proposed approach.

Original language	English
Title of host publication	2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publisher	IEEE
Pages	1846-1850
Number of pages	5
ISBN (Electronic)	9781479999880
DOIs	https://doi.org/10.1109/ICASSP.2016.7471996
Publication status	Published - 19 May 2016
Event	41st IEEE International Conference on Acoustics, Speech and Signal Processing 2016 - Shanghai International Convention Center, Shanghai, China Duration: 20 Mar 2016 → 25 Mar 2016

Conference

Conference	41st IEEE International Conference on Acoustics, Speech and Signal Processing 2016
Abbreviated title	ICASSP 2016
Country/Territory	China
City	Shanghai
Period	20/03/16 → 25/03/16

Keywords

Action recognition
BSIF
Low quality video
Mid-level representation
Texture features

ASJC Scopus subject areas

Software
Signal Processing
Electrical and Electronic Engineering

Access to Document

10.1109/ICASSP.2016.7471996

Cite this

@inproceedings{b9bf74a32a9048c9821bd38372095fcf,

title = "Spatio-temporal mid-level feature bank for action recognition in low quality video",

abstract = "It is a great challenge to perform high level recognition tasks on videos that are poor in quality. In this paper, we propose a new spatio-temporal mid-level (STEM) feature bank for recognizing human actions in low quality videos. The feature bank comprises of a trio of local spatio-temporal features, i.e. shape, motion and textures, which respectively encode structural, dynamic and statistical information in video. These features are encoded into mid-level representations and aggregated to construct STEM. Based on the recent binarized statistical image feature (BSIF), we also design a new spatiotemporal textural feature that extracts discriminately from 3D salient patches. Extensive experiments on the poor quality versions/subsets of the KTH and HMDB51 datasets demonstrate the effectiveness of the proposed approach.",

keywords = "Action recognition, BSIF, Low quality video, Mid-level representation, Texture features",

author = "Saimunur Rahman and John See",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE. Copyright: Copyright 2016 Elsevier B.V., All rights reserved.; 41st IEEE International Conference on Acoustics, Speech and Signal Processing 2016, ICASSP 2016 ; Conference date: 20-03-2016 Through 25-03-2016",

year = "2016",

month = may,

day = "19",

doi = "10.1109/ICASSP.2016.7471996",

language = "English",

pages = "1846--1850",

booktitle = "2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",

publisher = "IEEE",

address = "United States",

}

Rahman, S & See, J 2016, Spatio-temporal mid-level feature bank for action recognition in low quality video. in 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, pp. 1846-1850, 41st IEEE International Conference on Acoustics, Speech and Signal Processing 2016, Shanghai, China, 20/03/16. https://doi.org/10.1109/ICASSP.2016.7471996

TY - GEN

T1 - Spatio-temporal mid-level feature bank for action recognition in low quality video

AU - Rahman, Saimunur

AU - See, John

PY - 2016/5/19

Y1 - 2016/5/19

N2 - It is a great challenge to perform high level recognition tasks on videos that are poor in quality. In this paper, we propose a new spatio-temporal mid-level (STEM) feature bank for recognizing human actions in low quality videos. The feature bank comprises of a trio of local spatio-temporal features, i.e. shape, motion and textures, which respectively encode structural, dynamic and statistical information in video. These features are encoded into mid-level representations and aggregated to construct STEM. Based on the recent binarized statistical image feature (BSIF), we also design a new spatiotemporal textural feature that extracts discriminately from 3D salient patches. Extensive experiments on the poor quality versions/subsets of the KTH and HMDB51 datasets demonstrate the effectiveness of the proposed approach.

AB - It is a great challenge to perform high level recognition tasks on videos that are poor in quality. In this paper, we propose a new spatio-temporal mid-level (STEM) feature bank for recognizing human actions in low quality videos. The feature bank comprises of a trio of local spatio-temporal features, i.e. shape, motion and textures, which respectively encode structural, dynamic and statistical information in video. These features are encoded into mid-level representations and aggregated to construct STEM. Based on the recent binarized statistical image feature (BSIF), we also design a new spatiotemporal textural feature that extracts discriminately from 3D salient patches. Extensive experiments on the poor quality versions/subsets of the KTH and HMDB51 datasets demonstrate the effectiveness of the proposed approach.

KW - Action recognition

KW - BSIF

KW - Low quality video

KW - Mid-level representation

KW - Texture features

UR - https://www.scopus.com/pages/publications/84973301076

U2 - 10.1109/ICASSP.2016.7471996

DO - 10.1109/ICASSP.2016.7471996

M3 - Conference contribution

AN - SCOPUS:84973301076

SP - 1846

EP - 1850

BT - 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

PB - IEEE

T2 - 41st IEEE International Conference on Acoustics, Speech and Signal Processing 2016

Y2 - 20 March 2016 through 25 March 2016

ER -

Spatio-temporal mid-level feature bank for action recognition in low quality video

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this