Cascade method for image processing based people detection and counting

Zeyad Al-Zaydi, David Lorater Ndzi, David Adrian Sanders

Research output: Chapter in Book/Report/Conference proceedingChapter

39 Downloads (Pure)

Abstract

People detection is of great importance in video surveillance. Different approaches have been proposed to achieve accurate detection system. The main problem in people detection systems is that it must maintain a balance between the number of false detections and the number of missing people which limits the global detection results. In order to solve this problem and add robustness to detection, we propose a multiplexor and collector model composed of multiple independent detectors. This model is used to keep the true positive detections provided by a number of detectors and reduce the miss rate. In addition, a fusion model is proposed to check the robustness of the cascaded detection system. A pipeline techniques will also be used to avoid the increasing of detection time.
Original languageEnglish
Title of host publicationInt’l Conference Proceedings of International Conference on Image Processing, Production and Computer Science (ICIPCS'2016)
Subtitle of host publicationMarch 26-27, 2016 London (UK)
EditorsOsama Mohamed Mohamed Ahamed, Abhay Saxena
PublisherUniversal Researchers in Civil and Architecture Engineering
Pages30-36
Number of pages7
ISBN (Print)978-93-84422-71-4
DOIs
Publication statusPublished - 1 Mar 2016
Externally publishedYes

Keywords

  • people detection, counting, surveillance systems, image processing, computer vision

Fingerprint

Dive into the research topics of 'Cascade method for image processing based people detection and counting'. Together they form a unique fingerprint.

Cite this