Human activity recognition in low quality videos using spatio-temporal features

Human activity recognition (HAR) is one of the most intensively studied areas of computer vision in recent times. However, under real world conditions, especially when public infrastructure such as surveillance and web cameras are considered, current HAR techniques do not adapt to lower quality vide...

Full description

Saved in:

Bibliographic Details
Main Author:	Rahman, Saimunur
Format:	Thesis
Published:	2016
Subjects:	TK7800-8360 Electronics
Tags:	Add Tag No Tags, Be the first to tag this record!

id	my-mmu-ep.12830
record_format	uketd_dc
spelling	my-mmu-ep.128302024-08-16T03:34:40Z Human activity recognition in low quality videos using spatio-temporal features 2016-06 Rahman, Saimunur TK7800-8360 Electronics Human activity recognition (HAR) is one of the most intensively studied areas of computer vision in recent times. However, under real world conditions, especially when public infrastructure such as surveillance and web cameras are considered, current HAR techniques do not adapt to lower quality videos due to various challenges such as noise and lighting changes, motion blur, poor resolution and sampling. The objective of this research is to develop a framework and methods for human activity recognition using spatio-temporal information from low quality video. Overall, it can be observed that texture is an important visual feature cue for low quality video, and the robustness of shape and motion feature can be strengthened by using this. 2016-06 Thesis https://shdl.mmu.edu.my/12830/ http://erep.mmu.edu.my/ masters Multimedia University Faculty of Computing and Informatics (FCI) EREP ID: 12257
institution	Multimedia University
collection	MMU Institutional Repository
topic	TK7800-8360 Electronics
spellingShingle	TK7800-8360 Electronics Rahman, Saimunur Human activity recognition in low quality videos using spatio-temporal features
description	Human activity recognition (HAR) is one of the most intensively studied areas of computer vision in recent times. However, under real world conditions, especially when public infrastructure such as surveillance and web cameras are considered, current HAR techniques do not adapt to lower quality videos due to various challenges such as noise and lighting changes, motion blur, poor resolution and sampling. The objective of this research is to develop a framework and methods for human activity recognition using spatio-temporal information from low quality video. Overall, it can be observed that texture is an important visual feature cue for low quality video, and the robustness of shape and motion feature can be strengthened by using this.
format	Thesis
qualification_level	Master's degree
author	Rahman, Saimunur
author_facet	Rahman, Saimunur
author_sort	Rahman, Saimunur
title	Human activity recognition in low quality videos using spatio-temporal features
title_short	Human activity recognition in low quality videos using spatio-temporal features
title_full	Human activity recognition in low quality videos using spatio-temporal features
title_fullStr	Human activity recognition in low quality videos using spatio-temporal features
title_full_unstemmed	Human activity recognition in low quality videos using spatio-temporal features
title_sort	human activity recognition in low quality videos using spatio-temporal features
granting_institution	Multimedia University
granting_department	Faculty of Computing and Informatics (FCI)
publishDate	2016
_version_	1811768004748247040

Human activity recognition in low quality videos using spatio-temporal features

Similar Items