Biometrics, Security and Wrinkled Fingerprints

768 Words2 Pages

B. Wet and Wrinkled Finger Dataset
To test the working of algorithm wet as well as wrinkled (WWF) dataset is used. In Wet and Wrinkled Finger (WWF) database . Data from 30 people for all ten fingers using a multispectral fingerprint scanner from Lumidigm (Venus series) was collected. 300 fingers were treated as separate identities. Multispectral sensors were specially used as they were effective for application . They were designed to function when the fingers are wet with dripping water, and they can acquire an image when the finger is not in contact with the platen. This is possible because the multispectral sensor was able to acquire subsurface features as well as surface feature even under poor conditions; this contrasts with frustrated total Internal reaction sensors that require sufficient moisture along the ridges, air gaps in the valleys, and a clean dry platen. For each finger, the database contains two types of images: a pressed image and an air image. A pressed image is a regular scan of a finger that is pressed against the platen. The air image is an image of a fingertip that is not pressed on the sensor platen and lies just above it. The sensor produces a grayscale composite image from the multispectral signal, a raw RGB image to visually inspect the fingerprint, and a quality image. In total there were 3600 acquisitions because each of the 300 fingers has four modes (Dry-Air, Dry-Pressed, Wet-Air and Wet-Pressed) and each of the four modes has 3 repetitions for samples. Inspection of 300 air images revealed wrinkling in 185 images, and the corresponding fingers have been labeled as having exhibited wrinkling. To stimulate finger wrinkling, both hands of thirty subjects were soaked for 30 minutes in warm water main...

... middle of paper ...

...hniques such as correlation-based matching, minutiae-based matching, and pattern-based (or image-based) matching uses standard dataset for testing purpose. But Practically due to some physical changes in finger during verification ,system gets failed. Various fingerprint matching techniques do not authenticate wrinkled fingers. Thus error rate gets increased when matching is done between dry and wet-wrinkled fingers .Thus proposed system will extract features which will not change even after wrinkling. The proposed system will use minutiae based matching due to which error rate can be reduced. The Wet and Wrinkled Fingerprint (WWF) dataset is used to check the performance of proposed system. In this dataset there are wrinkled fingers due to wetness also some samples of dry fingers. Thus proposed matching algorithm will improve fingerprint recognition for wet fingers.

More about Biometrics, Security and Wrinkled Fingerprints

Open Document