Face Recognition Based on Local Fisher Features

摘要: To efficiently solve human face image recognition problem with an database, many techniques have been proposed. A key step in these is the extraction of features for indexing database and afterwards fulfilling tasks. Linear Discriminate Analysis(LDA) a statistic method classification. LDA filter global space local frequency. It squeezes all discriminant information into few basis vectors so that interpretation extracted becomes difficult. In this paper, we propose new idea to enhance performance recognition. We extract localized images by virtue wavelet transform. The simulation results suggest good classification ability our proposed system.