Human age estimation using Multi-Frequency Biologically Inspired Features (MF-BIF)

2D Gabor filters are sinusoidal waves of some frequency and orientation within a 2D Gaussian envelope and are used to model simple cells (S1) in primate visual cortex. Visual system in primates extracts information both in 2D spatial and frequency domains. In this paper, we use multiple spatial frequencies to model S1 cells in primates visual cortex. Thereafter, we use MAX and STD operators to extract Multi-Frequency Biologically Inspired Features (MF-BIF) for age estimation. We adopt hierarchical age estimation that consists of between-group classification followed by with-in group regression. This approach was evaluated on FG-NET aging dataset and results show that MF-BIF outperforms BIF estimation techniques.