Visual Tracking Based on Human Feature Extraction from Surveillance Video for Human Recognition

Main Article Content

Monisha G S
M.Hari Krishnan
Vetri Selvan M
G. Nirmala
Yogashree G S


A multimodal human identification system based on face and body recognition may be made available for effective biometric authentication. The outcomes are achieved by extracting facial recognition characteristics using several extraction techniques, including Eigen-face and Principle Component Analysis (PCA). Systems for authenticating people using their bodies and faces are implemented using artificial neural networks (ANN) and genetic optimization techniques as classifiers. Through feature fusion and scores fusion, the biometric systems for the human body and face are merged to create a single multimodal biometric system. Human bodies may be identified with astonishing accuracy and effectiveness thanks to the SDK for the Kinect sensor. To identify people, biometrics aims to mimic the pattern recognition process. In comparison to traditional authentication methods based on secrets and tokens, it is a more dependable and safe option. Human physiological and behavioral traits are used by biometric technologies to identify people automatically. These characteristics must fulfill many criteria, especially those that relate to universality, efficacy, and applicability.

Article Details

How to Cite
G S, M. ., Krishnan, M. ., Selvan M, V. ., Nirmala, G. ., & G S, Y. . (2023). Visual Tracking Based on Human Feature Extraction from Surveillance Video for Human Recognition. International Journal on Recent and Innovation Trends in Computing and Communication, 11(7), 133–141.


HaritaogluI,HarwoodD,DavisL(1999)Hydra:multiplepeopledetectionandtrackingusingsilhou-ettes. In: Proceedings of the international conference on image analysis and processing, 27–29 Sept.IEEEComputerSociety,Venice, pp280–285

CutlerR,DavisL(2000)Robustreal-timeperiodicmotiondetection,analysisandapplications.IEEETransPattern AnalMach Intell22(8):781–796



KaewTraKulPong P, Bowden R (2001) An improved adaptive background mixture model for realtimetrackingwithshadowdetection.In:Proceedingsof2ndEuropeanworkshoponadvancedvideobasedsurveillance systems, AVBS01. Sept 2001. Video based surveillance systems: computer vision anddistributedprocessing,KluwerAcademicPublishers,pp1–5

GrimsonWEL,StaufferC,RomanoR,LeeL(1998)Usingadaptivetrackingtoclassifyandmonitoractivities in a site. In: Proceedings of 1998 IEEE Computer Society conference on computer visionandpattern recognition(Cat.No.98CB36231).IEEEComputerSociety

Stauffer C, Grimson WEL (1998) Adaptive background mixture models for real-time tracking. In:CVPR1998


Bhavya K. R, & S. Pravinth Raja. (2023). Fruit Quality Prediction using Deep Learning Strategies for Agriculture. International Journal of Intelligent Systems and Applications in Engineering, 11(2s), 301–310. Retrieved from

ChenJ,ChenZ,ChiZ,FuH(2014)Facialexpressionrecognitionbasedonfacialcomponentsdetec-tion and HOG features. Scientific cooperations international workshops on electrical and computerengineeringsubfields,conductedbyKocUniversity,Turkey,pp64–69,duringtheperiod22–23

Köstinger M (2013) Efficient metric learning for real-world face recognition. Thesis Dissertation,InstituteforComputerGraphicsandVision,GrazUniversityofTechnology,Graz,Austria

Heisele B, Ho P, Poggio T (2001) Face recognition with support vector machines: global versuscomponent-basedapproach.In:ICCV

Shanthi, D. N. ., & J, S. . (2022). Social Network Based Privacy Data Optimization Using Ensemble Deep Learning Architectures. Research Journal of Computer Systems and Engineering, 3(1), 62–66. Retrieved from


Kobayashi T, Hidaka A, Kurita T (2008) Selection of histograms of oriented gradients features forpedestriandetection. In:ICONIP.Springer,pp598–607