Fingerprint Classification Based on Support Vector Machine

A fingerprint classification method using the fingerprint orientation information and Support Vector Machine (SVM) was proposed. Firstly, the reference point of fingerprint image was acquired, and fingerprint pattern area was located. Then the orientation features were extracted by using Gabor filter. Finally multiclass classifiers were constructed based on SVM and a new three stage classification approach was put forward. NIST-4 database is used to assess the proposed algorithms. The experimental results show our method can get good classification results.