Final published version
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - Image resolution impact analysis on pedestrian detection in smart cities surveillance
AU - Dinakaran, Ranjith
AU - Sexton, Graham
AU - Seker, Huseyin
AU - Bouridane, Ahmed
AU - Jiang, Richard
PY - 2017/10/17
Y1 - 2017/10/17
N2 - In the paradigm of smart cities 1 , video surveillance is becoming a widely-applied technology to help improve the quality of human life in the era of digital living, where pedestrian detection is one of the key components for people-centered smart cities applications including wellbeing, security, traffic guiding, and unmanned vehicles, etc. While so far most surveillance cameras are of low quality resolution for cost-saving reasons, the impact of the image resolution on detection accuracy becomes a concerned issue. Though lower resolution can save the cost as well as the processing time, it has not been clearly reported how the resolution can impact on the detection accuracy. In this paper, we investigate the limit of low-resolution cameras with regards to the accuracy of the pedestrian detection, and experimentally demonstrate its impact on the most widely-applied HOG-SVM pedestrian detector, which is a combination of Histogram of Oriented Gradient (HOG) and Support Vector Machine (SVM) for pedestrian detection. From our experiments, it is found that there is an optimal resolution to balance between speed and accuracy, while we show in our experiments that the resolution has apparent influence on both the accuracy and the computing time.
AB - In the paradigm of smart cities 1 , video surveillance is becoming a widely-applied technology to help improve the quality of human life in the era of digital living, where pedestrian detection is one of the key components for people-centered smart cities applications including wellbeing, security, traffic guiding, and unmanned vehicles, etc. While so far most surveillance cameras are of low quality resolution for cost-saving reasons, the impact of the image resolution on detection accuracy becomes a concerned issue. Though lower resolution can save the cost as well as the processing time, it has not been clearly reported how the resolution can impact on the detection accuracy. In this paper, we investigate the limit of low-resolution cameras with regards to the accuracy of the pedestrian detection, and experimentally demonstrate its impact on the most widely-applied HOG-SVM pedestrian detector, which is a combination of Histogram of Oriented Gradient (HOG) and Support Vector Machine (SVM) for pedestrian detection. From our experiments, it is found that there is an optimal resolution to balance between speed and accuracy, while we show in our experiments that the resolution has apparent influence on both the accuracy and the computing time.
KW - Histogram of oriented gradient (HOG)
KW - Image resolution
KW - Pedestrian detection
KW - Smart cities
KW - Support vector machine (SVM)
U2 - 10.1145/3109761.3109797
DO - 10.1145/3109761.3109797
M3 - Conference contribution/Paper
AN - SCOPUS:85048367687
T3 - ACM International Conference Proceeding Series
BT - Proceedings of the International Conference on Internet of Things and Machine Learning, IML 2017
A2 - Hamdan, Hani
A2 - Hidoussi, Faouzi
A2 - Boubiche, Djallel Eddine
PB - Association for Computing Machinery (ACM)
CY - New York
T2 - 1st International Conference on Internet of Things and Machine Learning, IML 2017
Y2 - 17 October 2017 through 18 October 2017
ER -