Hierarchical classification of Fingerprints using Discrete Wavelet Transform

이산 웨이블릿 변환을 이용한 지문의 계층적 분류

  • 권용호 (강원대학교 대학원 제어계측공학과) ;
  • 이정문 (강원대학교 전기전자공학부)
  • Published : 1999.12.31


An efficient method is developed for classifying fingerprint data based on 2-D discrete wavelet transform. Fingerprint data is first converted to a binary image. Then a multi-level 2-D wavelet transform is performed. Vertical and horizontal subbands of the transformed data show typical energy distribution patterns relevant to the fingerprint categories. The proposed method with moderate level of wavelet transform is successful in classifying fingerprints into 5 different types. Finer classification is possible by higher frequency subbands and closer analysis of energy distribution.