Minutia-based enhancement of fingerprint samples

Image enhancement is a common pre-processing step before the extraction of biometric features from a fingerprint sample. This can be essential especially for images of low image quality. An ideal fingerprint image enhancement should intend to improve the end-to-end biometric performance, i.e. the performance achieved on biometric features extracted from enhanced fingerprint samples. We use a model from Deep Learning for the task of image enhancement. This work's main contribution is a dedicated cost function which is optimized during training The cost function takes into account the biometric feature extraction. Our approach intends to improve the accuracy and reliability of the biometric feature extraction process: No feature should be missed and all features should be extracted as precise as possible. By doing so, the loss function forced the image enhancement to learn how to improve the suitability of a fingerprint sample for a biometric comparison process. The effectivity of the cost function was demonstrated for two different biometric feature extraction algorithms.