Font Size: a A A

Face Hiding Based On Detecting And Tracking

Posted on:2008-09-13Degree:MasterType:Thesis
Country:ChinaCandidate:X R ChenFull Text:PDF
GTID:2178360242460684Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
With the development of social economic,video survilliance is applied in manyoccasions,such as in traffic,bank,office,community, so many as domicile.It hasfiltered in many aspects of production and life.In broad application of videosurvilliance,it come forth a problem—personal privacy problem.With theimprovement of security and law consciousness,privacy protection of videosurvilliance arise many attention.A lot of foreign researchers have doneinterrelated works in this field,this work has very important academic significanceand pratical value of application.The main content of this dissertation is to build a face protection system onH.264 video coding in accordance with the process as face detection, face trackingand encryption.The system encrypt the face region in video which exposes people'sidentity so that we can protect the sensitive privacy information in picture.Thework do research on the key issues of each step. Main work and innovation are asfollows:(1) On the research of face detection, the dissertation achieves face detectionby using Adaboost in face detection which combines with a innovative Bayesclassifier validation.This method not only locate face exactly, but also reduce falsepositive rate on using Adaboost solely.Experiments proves the availability of thismethod.(2) On the research of face tracking, the dissertation solves the problem offace covered up using the integrated information of skin-color,size,predictablelocation,coefficient and histogram. The work do face tracking research on the keyissues of object coming and object disappearance and proposes a integrated methodof multi-face tracking. Experiments proves the availability of this method.(3) On the research of face encryption, the dissertation achieves encryptionand coding based on CABAC and RC coefficients.The work encryt the blocks ininterested face region with CABAC coding on H.264,which efficiently protectsface region.(4) The dissertation builds a face protection system based on face detection,tracking and face encryption.Experiments proves the availability of this method.
Keywords/Search Tags:Bayes classifier, Adaboost, Face tracking, H.264, Face encryption
PDF Full Text Request
Related items