Jianjiang Feng's Homepage

Home Research Publication Software Teaching

Latent fingerprint feature extraction Latent fingerprint feature extraction
Automatic feature extraction in latent fingerprints is extremely challenging due to poor quality of most latents, such as unclear ridge structures, overlapped lines and letters, and overlapped fingerprints.
Separating overlapped fingerprints Separating overlapped fingerprints
Overlapped fingerprints are not unusual in latent fingerprints taken from crime scenes. However, separating overlapped fingerprints into individual fingerprints is a very challenging problem for both existing automatic fingerprint recognition systems and human fingerprint experts. We proposed algorithms for separating overlapped fingerprints.
Latent fingerprint matching Latent fingerprint matching
Latent fingerprint identification is of critical importance for identifying suspects. Poor quality of ridge impressions, small finger area and large non-linear distortion are the main difficulties in latent fingerprint matching, compared to plain or rolled fingerprint matching.
Latent palmprint Latent palmprint matching
Latent palmprint matching is a challenging problem because latent prints lifted at crime scenes are of poor image quality, cover only a small area of the palm, have a complex background and many creases.
Altered fingerprint Fingerprint quality assessment
Fingerprint quality assessment is important for capturing high quality fingerprints and detecting uncooperative behaviors. We propose several algorithms to detect low quality fingerprints caused by different factors.
Fingerprint reconstruction Fingerprint reconstruction from minutiae
The compactness of minutiae representation has created an impression that minutiae does not contain sufficient information to allow the reconstruction of the original fingerprint image. This opinion is challenged by the proposed fingerprint reconstruction technique. The proposed technique also has other uses, such as reconstructing latent prints using minutiae marked by human experts, and improving the performance of fingerprint recognition systems with limitted storage space.
Ridge Matching Ridge skeleton matching
The most popular fingerprint representation is based on minutiae points, which is a compact and lossy representation of ridge skeleton, the Level 2 representation of fingerprints. Here we explore matching ridge skeletons directly for the purpose of completely utilizing the discriminating power of Level 2 features in fingerprints.
Minutia descriptor Minutia descriptors
Establishing minutiae correspondence between two fingerprints is difficult due to unknown alignment, nonlinear deformation, noise, and occlusion. Designing robust minutia descriptors is a way to facilitate this problem.
Synthetic palmprint Statistical modeling of friction ridge patterns
Statistical modeling of friction ridge patterns is a fundamental problem and can be used in both synthesis and analysis of friction ridge patterns.