Versatile loitering detection based on non-verbal cues using dense trajectory descriptors.
Saved in:
| Title: | Versatile loitering detection based on non-verbal cues using dense trajectory descriptors. |
|---|---|
| Authors: | Arivazhagan S1, Newlin Shebiah R1 newlinshebiah@yahoo.co.in |
| Source: | Multimedia Tools & Applications. Apr2019, Vol. 78 Issue 8, p10933-10963. 31p. |
| Subjects: | Loitering, Nonverbal cues, Criminal investigation, Support vector machines, Euclidean distance |
| Abstract: | Loitering analytics is widely explored nowadays since it anticipates crimes by identifying suspicious behavioural pattern displayed by offenders and reacts to the circumstances without delay. The challenge in loitering detection lies in effective discrimination of the behavioural pattern of offensive loiter and innocuous loiter. In this paper, certain embodiments were proposed with dense trajectory features representing human gait parameters since, many of the Personality traits are manifested in it. Here, frame differencing of effectively represented frames by wavelet transform is used for moving blob detection. With the prior model developed from benchmark datasets using co-occurrence features, the motion blobs were classified as pedestrian or other moving objects using SVM classifier. Short term biometric features like the clothing colour and texture are successfully used to track the person in successive frames. Since the perceiver can make judgment on the target within 10 s under unacquainted condition, here short sequences of frames representing 10 s of video is used for processing. For the short sequence, data association matrix relating the frame number and the associated pedestrians in each frame based on minimum Euclidean distance is proposed. Missing tracks due to occlusion can be effectively handled by a completely unsupervised system using the proposed data association matrix. From the developed matrix the person staying in the Region of Interest for a long duration is identified and behavioural cues displayed by the person are extracted. Here, the spatio- temporal features extracted from Dense Trajectories and Motion Boundary Descriptors from the pre-learned model is used to characterize the person as loiter or not. To evaluate the performance of the proposed method, PETS 2006, PETS 2007 and PETS 2016 datasets were used and the experiments show promising results comparable with the state of art techniques. [ABSTRACT FROM AUTHOR] |
| Copyright of Multimedia Tools & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| FullText | Links: – Type: pdflink Text: Availability: 1 |
|---|---|
| Header | DbId: egs DbLabel: Engineering Source An: 136338608 AccessLevel: 6 PubType: Academic Journal PubTypeId: academicJournal PreciseRelevancyScore: 0 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Versatile loitering detection based on non-verbal cues using dense trajectory descriptors. – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Arivazhagan+S%22">Arivazhagan S</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Newlin+Shebiah+R%22">Newlin Shebiah R</searchLink><relatesTo>1</relatesTo><i> newlinshebiah@yahoo.co.in</i> – Name: TitleSource Label: Source Group: Src Data: <searchLink fieldCode="JN" term="%22Multimedia+Tools+%26+Applications%22">Multimedia Tools & Applications</searchLink>. Apr2019, Vol. 78 Issue 8, p10933-10963. 31p. – Name: Subject Label: Subjects Group: Su Data: <searchLink fieldCode="DE" term="%22Loitering%22">Loitering</searchLink><br /><searchLink fieldCode="DE" term="%22Nonverbal+cues%22">Nonverbal cues</searchLink><br /><searchLink fieldCode="DE" term="%22Criminal+investigation%22">Criminal investigation</searchLink><br /><searchLink fieldCode="DE" term="%22Support+vector+machines%22">Support vector machines</searchLink><br /><searchLink fieldCode="DE" term="%22Euclidean+distance%22">Euclidean distance</searchLink> – Name: Abstract Label: Abstract Group: Ab Data: Loitering analytics is widely explored nowadays since it anticipates crimes by identifying suspicious behavioural pattern displayed by offenders and reacts to the circumstances without delay. The challenge in loitering detection lies in effective discrimination of the behavioural pattern of offensive loiter and innocuous loiter. In this paper, certain embodiments were proposed with dense trajectory features representing human gait parameters since, many of the Personality traits are manifested in it. Here, frame differencing of effectively represented frames by wavelet transform is used for moving blob detection. With the prior model developed from benchmark datasets using co-occurrence features, the motion blobs were classified as pedestrian or other moving objects using SVM classifier. Short term biometric features like the clothing colour and texture are successfully used to track the person in successive frames. Since the perceiver can make judgment on the target within 10 s under unacquainted condition, here short sequences of frames representing 10 s of video is used for processing. For the short sequence, data association matrix relating the frame number and the associated pedestrians in each frame based on minimum Euclidean distance is proposed. Missing tracks due to occlusion can be effectively handled by a completely unsupervised system using the proposed data association matrix. From the developed matrix the person staying in the Region of Interest for a long duration is identified and behavioural cues displayed by the person are extracted. Here, the spatio- temporal features extracted from Dense Trajectories and Motion Boundary Descriptors from the pre-learned model is used to characterize the person as loiter or not. To evaluate the performance of the proposed method, PETS 2006, PETS 2007 and PETS 2016 datasets were used and the experiments show promising results comparable with the state of art techniques. [ABSTRACT FROM AUTHOR] – Name: AbstractSuppliedCopyright Label: Group: Ab Data: <i>Copyright of Multimedia Tools & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.) |
| PLink | https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=136338608 |
| RecordInfo | BibRecord: BibEntity: Identifiers: – Type: doi Value: 10.1007/s11042-018-6618-9 Languages: – Code: eng Text: English PhysicalDescription: Pagination: PageCount: 31 StartPage: 10933 Subjects: – SubjectFull: Loitering Type: general – SubjectFull: Nonverbal cues Type: general – SubjectFull: Criminal investigation Type: general – SubjectFull: Support vector machines Type: general – SubjectFull: Euclidean distance Type: general Titles: – TitleFull: Versatile loitering detection based on non-verbal cues using dense trajectory descriptors. Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Arivazhagan S – PersonEntity: Name: NameFull: Newlin Shebiah R IsPartOfRelationships: – BibEntity: Dates: – D: 15 M: 04 Text: Apr2019 Type: published Y: 2019 Identifiers: – Type: issn-print Value: 13807501 Numbering: – Type: volume Value: 78 – Type: issue Value: 8 Titles: – TitleFull: Multimedia Tools & Applications Type: main |
| ResultId | 1 |
