Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/319710
Title: Performance Enhancement of Automated Video Surveillance Systems
Researcher: Manju D
Guide(s): Radha V
Keywords: Engineering and Technology
Computer Science
University: Avinashilingam Deemed University For Women
Completed Date: 2021
Abstract: Considering the security concerns, it has become necessary to enforce a sophisticated Video Surveillance System (VSS) in order to enable the human agent to control public places. Surveillance cameras are common and they produce an incredible amount of data every day. In particular, surveillance cameras are mounted in large areas to capture any moment of the event, for offline video analysis and evaluation of suspicious events in places such as the departmental shops, complex and the market. But when this is done manually, the analysis and detection of abnormal activities in video surveillance system becomes complicated. So, using video in machine understanding is an important research topic. One of the interesting topics in video surveillance system is activity understanding. In this research work, various approaches are proposed to detect and confirm the abnormal activities in the video surveillance data. newlineThe objectives of this research work are to identify the frequent actions in the video surveillance data with less execution time and memory consumption, effective human activity prediction at night and rainy time, enable recognition of human faces with any view and angle range, and to estimate the age and gender of unknown person, newlineIn the first phase of the research work, Frequent Pattern-growth (FP-growth) is introduced to determine the frequent actions in the video surveillance data. Initially, spatial information, size and motion correlation among objects in the video frames are collected and then the objects are tracked by using partial filter method. The identified and tracked objects are converted into complex symbolic sequence and the frequent pattern is found from the complex symbolic sequence by using FP-Tree. The frequent patterns are taken as normal activities and the remaining patterns are treated as abnormal activities. The whole process is named as Spatio-Temporal Frequent Object Mining (STFOM). newlineThe STFOM prediction does not work well during rain and night times. Under rainy conditions and with ni
Pagination: 152
URI: http://hdl.handle.net/10603/319710
Appears in Departments:Department of Computer Science

Files in This Item:
File Description SizeFormat 
01_title.pdfAttached File118.21 kBAdobe PDFView/Open
02_certificate.pdf408.1 kBAdobe PDFView/Open
03_acknowledgement.pdf87.89 kBAdobe PDFView/Open
04-contents.pdf103.69 kBAdobe PDFView/Open
05_list of tables, figures, abbreviations and symbols.pdf291.67 kBAdobe PDFView/Open
06_chapter 1.pdf482.91 kBAdobe PDFView/Open
07_chapter 2.pdf354.17 kBAdobe PDFView/Open
08_chapter 3.pdf235.9 kBAdobe PDFView/Open
09_chapter 4.pdf483.96 kBAdobe PDFView/Open
10_chapter 5.pdf492.36 kBAdobe PDFView/Open
11_chapter 6.pdf487.24 kBAdobe PDFView/Open
12_chapter 7.pdf353.51 kBAdobe PDFView/Open
13_chapter 8.pdf163.36 kBAdobe PDFView/Open
14_bibliography.pdf183.12 kBAdobe PDFView/Open
80_recommendation.pdf54.56 kBAdobe PDFView/Open


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: