Face Recognition Technology |
||||
| HOME | SOURCE CODE | SOFTWARE INFO | SUPPORT | CONTACT US |
Statistical Moments for Pattern RecognitionDownload now Matlab source code Requirements: Matlab, Matlab Image Processing Toolbox. Moment based feature descriptors have evolved into a powerful tool for image analysis applications. Geometric moments present a low computational cost, but are highly sensitive to noise. Furthermore reconstruction is extremely difficult. Although not invariant under rotation, Hu's invariants that are derived from geometric moments present invariance under linear transformations. Complex moments provide with additional invariant descriptors, but present the same problems regarding noise and reconstruction. Moments of orthogonal polynomial basis were proposed by Teague. They have proven less sensitive to noise, are natively invariant to linear transformations and can be effectively used for image reconstruction. Computational complexity, however, becomes a major issue, and real-time implementation in software has not been reported. Moments of discrete orthogonal basis have been proposed recently. They are fast to implement, present adequate noise tolerance and very accurate image reconstruction. Their major drawback is the lack of invariance under transformation. Image normalization should be used prior to moment extraction for applications requiring invariance. We have developed a simple and efficient technique for face recognition that combines:
Each feature vector in fact is not discriminative for identification and only using them all at once with appropriate weights it is possible to reach an excellent recognition rate. Index Terms: Matlab, source, code, face, recognition, statistical, moments, moment, invariant, Hu, centralised, Legendre. Release 1.0 Date 2008.07.15 Major features:
|
Face Recognition . It Luigi Rosa mobile +39 3207214179 luigi.rosa@tiscali.it http://www.advancedsourcecode.com |