Skip to main content

Table 5 Available latent fingerprint datasets

From: An investigation of latent fingerprinting techniques

Dataset Description
NISD27 (, 2019) 258 samples of grayscale fingerprint images. Includes both 500 pixels per inch (PPI) and 1000 PPI samples. Manually annotated features are also available for sample images. Can be used for rolled fingerprint matching.
WVU latent databases (, 2019) Collection of 449 images. Contains exemplars of 500 and 1000 PPI marked features that are available.
The database can be used for latent to rolled fingerprint matching.
FVC2004 databases (, 2019) Collection of 1440 impressions. The database is constructed using 120 fingers with 12 impressions per finger.
DB1 and DB2 were collected using optical sensors
DB3 collected using thermal sweeping sensor
DB4 collected using synthetic fingerprint generation sensors
IIIT latent fingerprint database (, 2019) The database is a collection of 15 subjects (for each subject, there are 10 fingerprints). Grayscale images are scanned using a 500-PPI scanner. The size of each image is 4752 × 3168 pixels
IIIT Simultaneous Latent Fingerprint (SLF) database (, 2019) The database contains a simultaneous fingerprint of 15 subjects. Fingerprint images are obtained using the black powder technique.
IIITD Multi-surface Latent Fingerprint database (IIITD-MSLFD) (, 2019) Consists of 551 latent fingerprints samples. Includes 500 DPI samples. Samples of 51 subjects are captured
Eight different surfaces are used for capturing fingerprints (e.g. Ceramic mug, plate, steel glass, book cover, etc.)
IIITD Multisensor Optical and Latent Fingerprint database (Sankaran et al., 2015) The database contains 19,200 fingerprint samples. One-hundred subjects were used for the construction of the database.
Methods like CrossMatch L-Scan Patrol, Secugen Hamster, etc. are used
Tsinghua Latent Overlapped Fingerprint database (, 2019) Consists of 12 plain fingerprints and 100 latent fingerprints which are overlapped
Optical fingerprint scanners are used to capture the dataset
Includes 500 PPI samples
ELFT-EFS Public Challenge database(, 2019) The database contains 1100 images. Includes both 500 pixels per inch (PPI) and 1000 PPI samples. level 1, level 2, as well as level 3 features, can be extracted using this database