A novel face recognition model for fighting against human trafficking in surveillance videos and rescuing victims

The trade of humans aimed at the intention of forced labor, sexual slavery, commercial sexual slavery, or else others is termed human trafficking (HT). It happens inside a nation or else trans-nationality. Victims of trafficking may be of whichever age and gender. Herein, the face recognition (FR) d...

Full description

Saved in:
Bibliographic Details
Published in:Soft computing (Berlin, Germany) Germany), 2023-09, Vol.27 (18), p.13165-13180
Main Authors: Karpagam, M., Jeyavathana, R. Beaulah, Chinnappan, Sathiya Kumar, Kanimozhi, K. V., Sambath, M.
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Artificial Intelligence
Biometrics
Cameras
Computational Intelligence
Control
Engineering
Face recognition
Feature extraction
Focus
Genetic algorithms
Image acquisition
Image contrast
Image enhancement
Mathematical Logic and Foundations
Mechatronics
Motion simulation
Neural networks
Optical flow (image analysis)
Robotics
Surveillance
Video
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The trade of humans aimed at the intention of forced labor, sexual slavery, commercial sexual slavery, or else others is termed human trafficking (HT). It happens inside a nation or else trans-nationality. Victims of trafficking may be of whichever age and gender. Herein, the face recognition (FR) design to fight against HT is propounded centred on surveillance videos aimed at rescuing the victims. Initially, the missing persons’ image is gathered as of the police department and then undergoes pre-processing. In the pre-processing phase, filtering (Gaussian filter (GF)) is executed to eradicate the noise existent in the image, and contrast enhancement (IMF-CLAHE) is done to acquire the contrast-enhanced image (EI). Next, in the image, the face is identified utilizing Viola–Jones algorithm (VJA). Then, feature extraction (FE) occurs; after that dimensionality reduction utilizing P7SKLDA is executed. The dimensionality reduced features are then inputted into the MDBN classifier to recognize the person. After performing FR on the inputted data, FR is executed on the surveillance videos by transforming the videos to frames. Lastly, motion estimation utilizing the Horn–Schunck Optical flow (HSOF) methodology is utilized to enumerate the relative motion betwixt surveillance videos’ frames aimed at detecting the location of the missing one’s trafficking. Lastly, the proposed design’s performance is analogized with the existent methodologies to deduce this methodology’s superiority.
ISSN:1432-7643
1433-7479
DOI:10.1007/s00500-022-06931-1