Detecting method for reference point of fingerprint picture

Abstract

Abstract PURPOSE:To obtain a reference point by performing the artificial Fourier transform a detect the distribution state of picture elements within a small area after separating this small area, and therefore detecting roughtly a cetner point of the entire picture to scan the small area around the detected center point through comparison among areas. CONSTITUTION:A fingerprint is detected by a fingerprint picture extracting mechanism 1, and the picture data supplied to a TV camera part 2 is binary coded by a binary coding part 3 to be temporarily stored to a memory 4. Then the binary coded data is converted through a small area of each part. Then a feature extracting part 6 obtains a power spectrum ¦F(omega)¦, and a CPU5 compares the power spectra to extract a center point of a small area. In this case, an artifical Fourier transform is employed. In case a center point exists, the power spectrum ¦F(omega)¦ corresponding to the frequency omega corresponding to the cycle of a picture which is expanded radially has its peak value. Based on this peak value, the center point of a small area having the maximum power spectrum is decided as the center point of the fingerprint picture.

Inventors

• 👤 Tomoyuki Kiyosue
• 👤 悌之 清末
• 👤 Masahiko Hase
• 👤 雅彦 長谷
• 👤 Hiroyuki Hoshino
• 👤 星野　坦之
• 👤 Akihiro Shimizu
• 👤 明宏 清水

Bibliographic Details