One-shot facial feature extraction based on Gauss-Laguerre filter

This paper explores application of a single Gauss-Laguerre (GL) filter for facial feature extraction in the task of facial recognition. Such a one-shot procedure is a robust form of feature extraction performed on facial images without intensive pre-processing algorithms, while maintaining a relative recognition accuracy using the extracted features. The GL filters can be tuned specifically for each individual facial feature or for the entire face to extract useful information. This approach demonstrates a 78.12% and 92.5% recognition rate by applying a single GL filter on the FERET and ORL databases respectively. Unlike traditional methods of filter-based feature extraction that require a bank of filters to handle minor rotation and occlusion due to facial hair, GL requires only one application of an appropriately tuned filter.