A Spiking Neural Network Based Cortex-Like Mechanism and Application to Facial Expression Recognition

Figure 1

Framework of the entire facial expression recognition system. From the raw input to the final output, with each layer’s illustration included, which have shown that the combination of selectiveness, determinativeness and invariance is built up gradually across several stages of facial information processing. The preprocessing part includes the detection of facial regions(eye, mouth, etc.), illumination normalization, retina level is also responsible for edge detector for enhancing the high contrast of the image. The second level functions like Gabor filter, which send the output to the perceptual level for extracting features which are robust for selectivity and invariance, then after being grouped and classified, the category level gives the output results (best view in color) [22].