Pagrindinių komponentų analizės metodas žmogaus veidui atpažinti
| Year | Start Page | End Page |
|---|---|---|
2007 | 252 | 259 |
Darbe nagrinėjamas Pagrindinių komponentų analizavimo metodas. Aptariamas žmonių veidų atpažinimo algoritmas naudoant pagrindinių komponentų analizavimo metodą, yra aptarti šio metodo privalumai ir trūkumai. Pateikti atpažinimo eksperimento rezultatai.
Principle components analysis provides a theretically sound way of determining the underlying problem that explain some observation. hus PCA is perfectly suited to the problem of face recognition werw we need to explain hogh-dimensional observations with as few variables as possible. This decreases the computational complexity of the face recognition task, and scales each variabl according to it's relative importance in determining identity. Eigenfaces can be used to recognize and locate faces. Through the use of the Mahalanobis distance and assumption of gaussian distribution it is possible to determine the certainly that a face recognition is correct, and that the ectracted region of the scene actually corresponds to a face.