We investigate body soft biometrics capabilities to perform pruning of a hard biometrics database improving both retrieval speed and accuracy. Our pre-classification scheme based on anthropometricmeasures is elaborated on a large scale medical dataset to guarantee statistical meaning of the results, and tested in conjunction with a face recognition algorithm. Our assumptions are verified building and testing our system on a chimera dataset. We
clearly identify the trade off among pruning, accuracy, and mensuration error of an anthropomeasure based system. Our results show that even in the worst case of ±10% error magnitude in the anthropometric measures, our pruning scheme improves the accuracy performances guaranteeing a speedup of 2×factor.