Compression neural networks for feature extraction: Application to human recognition from ear images