Font Size: a A A

Researches On Face Detection Technology

Posted on:2008-10-31Degree:MasterType:Thesis
Country:ChinaCandidate:L TianFull Text:PDF
GTID:2178360212476520Subject:Electrical theory and new technology
Abstract/Summary:PDF Full Text Request
Face detection problem is firstly proposed as a key step of face recognition system. And with the extensive development of face recognition and advancement of developing practical system, face detect is developing into a independent research field. Up to now, the research of face detection algorithm is rather thorough; however, the multi-pose face detection in complex background is still a difficult problem. In addition, meeting the requirement of both speed and precision in a face detection system is another key issue. In order to solve the above difficulties, we have conducted the research on face detection in the following three ways, after introducing the current status of face detection research:1) Cascading algorithm based profile face detection using nonlinear dimension reduction.We propose a new profile face detection algorithm integrating nonlinear feature extraction, SVM classifiers and cascading strategy. Firstly, we select LE as the nonlinear feature extraction algorithm after comparing the current...
Keywords/Search Tags:face detection, nonlinear SVM, histogram-based statistical structure, clustering, binary decision tree, Adaboost algorithm
PDF Full Text Request
Related items