Illumination Planning for Object Recognition Using Parametric Eigenspaces
IEEE Transactions on Pattern Analysis and Machine Intelligence
Hi-index | 0.00 |
A dictionary registration process of subspace face recognition for realizing strong tolerances to lighting fluctuations is described in this paper. Camera devices have been added to handy phones and enable us to do biometric certification using face images. However, face images taken by handy phones are easily fluctuated by lighting conditions. Our proposed method is based on subspace face recognition which can realize robustness against fluctuations. Proposed method can create a subspace with robustness against lighting fluctuations virtually, using the lighting canonical space which contains various types of lighting elements. Not only equal-error-rate of verification but also precision of identification can be improved by proposed method, from 48.7% up to 91.9%.