An efficient framework for visual target tracking detection and action recognition

dc.contributor.guideRhymend Uthariaraj V and John Prakash S
dc.coverage.spatialAn efficient framework for visual target tracking detection and action recognition
dc.creator.researcherDeepika Roselind J
dc.date.accessioned2023-11-01T09:12:59Z
dc.date.available2023-11-01T09:12:59Z
dc.date.awarded2023
dc.date.completed2023
dc.date.registered
dc.description.abstractThis research presents a novel comprehensive framework for visual target tracking, detection, and action recognition to overcome some existing issues such as feature misalignment, occlusion handling, scale variations, complex backgrounds, and network issues. The first contribution of the research work is the development of an effective technique for detecting multi-scale and arbitrarily oriented objects. A revised single-stage detector is proposed for localizing and detecting oriented and multi-scale objects. The proposed detector uses an enhanced feature refinement module that reconstructs an enhanced feature map for accurate localization and detection of oriented and multi-scale targets. The second contribution is the proposal of an efficient algorithm for detecting and tracking visual targets in dense environments. The novel tracking system is modelled based on a correlation filtering-based approach. The proposed tracker partitions a target based on its size and shape to learn the discriminative features in each partition. Based on the correlation output (APCE value), the occlusion status is mapped for each partition. The proposed tracker includes a redetection mechanism to improve the efficiency of the tracker and to reduce its computation complexity. newline
dc.description.note
dc.format.accompanyingmaterialNone
dc.format.dimensions21 cm
dc.format.extentxx, 196
dc.identifier.urihttp://hdl.handle.net/10603/522234
dc.languageEnglish
dc.publisher.institutionFaculty of Information and Communication Engineering
dc.publisher.placeChennai
dc.publisher.universityAnna University
dc.relationp. 175-195
dc.rightsuniversity
dc.source.universityUniversity
dc.subject.keywordAPCE value
dc.subject.keywordComputer Science
dc.subject.keywordComputer Science Information Systems
dc.subject.keywordEngineering and Technology
dc.subject.keywordFramework
dc.subject.keywordProposed tracker
dc.titleAn efficient framework for visual target tracking detection and action recognition
dc.title.alternative
dc.type.degreePh.D.

Files

Original bundle

Now showing 1 - 5 of 12
Loading...
Thumbnail Image
Name:
01_title.pdf
Size:
143.66 KB
Format:
Adobe Portable Document Format
Description:
Attached File
Loading...
Thumbnail Image
Name:
02_prelim_pages.pdf
Size:
1.72 MB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
03_content.pdf
Size:
164.6 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
04_abstract.pdf
Size:
108.83 KB
Format:
Adobe Portable Document Format
Loading...
Thumbnail Image
Name:
05_chapter 1.pdf
Size:
154.44 KB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.79 KB
Format:
Plain Text
Description: