Biometric characterization systems are generally used in safety-related applications because they allow the identification or verification of individuals based on human body traits. In recent years hand veins have become an attractive biometric trait due to their advantages compared with other classical biometric traits (i.e., fingerprints, iris, face). However, due to the number of possible architectures for feature extraction and individual identification, different combinations between such methods should be evaluated to give a baseline for further vein biometrics development. This work presents a comparative analysis for individual identification based on hand-vein biometrics, which combines four feature extraction techniques and three classic machine learning techniques using two main types of images. The results show the reliability of some combinations for hand-vein biometric identification achieving accuracy levels above 98% and an Equal Error Rate under 3.2%.