Output details
13 - Electrical and Electronic Engineering, Metallurgy and Materials
Newcastle University
Multi-linear neighborhood preserving projection for face recognition
This research uses algorithms to work directly on multidimensional face images to extract features for biometric face recognition and classification. The work has introduced new mathematical theory and is operable in both unsupervised/supervised modes. Comparative evaluations with published work show that this approach solves the high ‘curse of dimensionality’ problem and reduces the high computational demand. The work lends itself to all high speed image recognition and retrieval applications and the initial stages of the research were a driver for the successful application to the Royal Society for an International Scientific Seminar for India and UK ISS/India/JEB.