Cloud-based scalable object detection and classification in video streams

Muhammad Usman Yaseen, Ashiq Anjum, Omer Rana, Richard Hill

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Due to the recent advances in cameras, cell phones and camcorders, particularly the resolution at which they can record an image/video, large amounts of data are generated daily. This video data is often so large that manually inspecting it for useful content can be time consuming and error prone, thereby it requires automated analysis to extract useful information and metadata. Existing video analysis systems lack automation, scalability and operate under a supervised learning domain, requiring substantial amounts of labelled data and training time. We present a cloud-based, automated video analysis system to process large numbers of video streams, where the underlying infrastructure is able to scale based on the number and size of the stream(s) being considered. The system automates the video analysis process and reduces manual intervention. An operator using this system only specifies which object of interest is to be located from the video streams. Video streams are then automatically fetched from the cloud storage and analysed in an unsupervised way. The proposed system was able to locate and classify an object of interest from one month of recorded video streams comprising 175 GB in size on a 15 node cloud in 6.52 h. The GPU powered infrastructure took 3 h to accomplish the same task. Occupancy of GPU resources in cloud is optimized and data transfer between CPU and GPU is minimized to achieve high performance. The scalability of the system is demonstrated along with a classification accuracy of 95%.
Original languageEnglish
Pages (from-to)286-298
Number of pages13
JournalFuture Generation Computer Systems
Volume80
Early online date7 Feb 2017
DOIs
Publication statusPublished - Mar 2018
Externally publishedYes

Fingerprint

Scalability
Supervised learning
Video cameras
Data transfer
Metadata
Program processors
Automation
Cameras
Object detection
Graphics processing unit

Cite this

Yaseen, Muhammad Usman ; Anjum, Ashiq ; Rana, Omer ; Hill, Richard. / Cloud-based scalable object detection and classification in video streams. In: Future Generation Computer Systems. 2018 ; Vol. 80. pp. 286-298.
@article{cc50f8ab5ab04ba988a63b2b717c9ea1,
title = "Cloud-based scalable object detection and classification in video streams",
abstract = "Due to the recent advances in cameras, cell phones and camcorders, particularly the resolution at which they can record an image/video, large amounts of data are generated daily. This video data is often so large that manually inspecting it for useful content can be time consuming and error prone, thereby it requires automated analysis to extract useful information and metadata. Existing video analysis systems lack automation, scalability and operate under a supervised learning domain, requiring substantial amounts of labelled data and training time. We present a cloud-based, automated video analysis system to process large numbers of video streams, where the underlying infrastructure is able to scale based on the number and size of the stream(s) being considered. The system automates the video analysis process and reduces manual intervention. An operator using this system only specifies which object of interest is to be located from the video streams. Video streams are then automatically fetched from the cloud storage and analysed in an unsupervised way. The proposed system was able to locate and classify an object of interest from one month of recorded video streams comprising 175 GB in size on a 15 node cloud in 6.52 h. The GPU powered infrastructure took 3 h to accomplish the same task. Occupancy of GPU resources in cloud is optimized and data transfer between CPU and GPU is minimized to achieve high performance. The scalability of the system is demonstrated along with a classification accuracy of 95{\%}.",
keywords = "Unsupervised object classification, Cloud computing, GPUs, High performance video analytics",
author = "Yaseen, {Muhammad Usman} and Ashiq Anjum and Omer Rana and Richard Hill",
year = "2018",
month = "3",
doi = "10.1016/j.future.2017.02.003",
language = "English",
volume = "80",
pages = "286--298",
journal = "Future Generation Computer Systems",
issn = "0167-739X",
publisher = "Elsevier",

}

Cloud-based scalable object detection and classification in video streams. / Yaseen, Muhammad Usman; Anjum, Ashiq; Rana, Omer; Hill, Richard.

In: Future Generation Computer Systems, Vol. 80, 03.2018, p. 286-298.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Cloud-based scalable object detection and classification in video streams

AU - Yaseen, Muhammad Usman

AU - Anjum, Ashiq

AU - Rana, Omer

AU - Hill, Richard

PY - 2018/3

Y1 - 2018/3

N2 - Due to the recent advances in cameras, cell phones and camcorders, particularly the resolution at which they can record an image/video, large amounts of data are generated daily. This video data is often so large that manually inspecting it for useful content can be time consuming and error prone, thereby it requires automated analysis to extract useful information and metadata. Existing video analysis systems lack automation, scalability and operate under a supervised learning domain, requiring substantial amounts of labelled data and training time. We present a cloud-based, automated video analysis system to process large numbers of video streams, where the underlying infrastructure is able to scale based on the number and size of the stream(s) being considered. The system automates the video analysis process and reduces manual intervention. An operator using this system only specifies which object of interest is to be located from the video streams. Video streams are then automatically fetched from the cloud storage and analysed in an unsupervised way. The proposed system was able to locate and classify an object of interest from one month of recorded video streams comprising 175 GB in size on a 15 node cloud in 6.52 h. The GPU powered infrastructure took 3 h to accomplish the same task. Occupancy of GPU resources in cloud is optimized and data transfer between CPU and GPU is minimized to achieve high performance. The scalability of the system is demonstrated along with a classification accuracy of 95%.

AB - Due to the recent advances in cameras, cell phones and camcorders, particularly the resolution at which they can record an image/video, large amounts of data are generated daily. This video data is often so large that manually inspecting it for useful content can be time consuming and error prone, thereby it requires automated analysis to extract useful information and metadata. Existing video analysis systems lack automation, scalability and operate under a supervised learning domain, requiring substantial amounts of labelled data and training time. We present a cloud-based, automated video analysis system to process large numbers of video streams, where the underlying infrastructure is able to scale based on the number and size of the stream(s) being considered. The system automates the video analysis process and reduces manual intervention. An operator using this system only specifies which object of interest is to be located from the video streams. Video streams are then automatically fetched from the cloud storage and analysed in an unsupervised way. The proposed system was able to locate and classify an object of interest from one month of recorded video streams comprising 175 GB in size on a 15 node cloud in 6.52 h. The GPU powered infrastructure took 3 h to accomplish the same task. Occupancy of GPU resources in cloud is optimized and data transfer between CPU and GPU is minimized to achieve high performance. The scalability of the system is demonstrated along with a classification accuracy of 95%.

KW - Unsupervised object classification

KW - Cloud computing

KW - GPUs

KW - High performance video analytics

UR - http://www.scopus.com/inward/record.url?scp=85013031070&partnerID=8YFLogxK

U2 - 10.1016/j.future.2017.02.003

DO - 10.1016/j.future.2017.02.003

M3 - Article

VL - 80

SP - 286

EP - 298

JO - Future Generation Computer Systems

JF - Future Generation Computer Systems

SN - 0167-739X

ER -