Fingerprint Cards patent filing for biometric fingerprint image enhancement published in U.S.
A patent application for a biometric fingerprint image enhancement system from Fingerprint Cards has been published in the U.S., with the promise of improved matching performance through reducing the distortion that can occur when capturing fingerprint images on electronic devices.
Spotted and tweeted by Stefan J. Jensen, the patent filing describes combining different coefficients from a pair of fingerprint image data sets produced by low pass filters to produce an enhanced image. The filing for a ‘Method and system for fingerprint image enhancement” refers to the use of biometric systems in consumer electronic devices for increased user security and convenience.
The traditional method of capturing fingerprints with a pixel matrix typically distorts raw fingerprint images in different ways, according to the application, such as blurring, non-uniformity, or the inclusion of noise, possibly “from gain stages in analog circuitry of the fingerprint sensor.” Blur can be caused by variations in the electric field between the finger and the sensor area. De-blurring is generally performed by inverting these effects, but FPC writes that that approach may create artifacts elsewhere in the image.
By applying at least two low pass filters and two weights to the raw fingerprint image data separately, at least two filtered sets of fingerprint data can be obtained and combined, which according to the patent application could deal with distortions like non-uniformity in the image to improve biometric fingerprint recognition performance.
The filters may be Gaussian, and the sensor may be capacitive, but neither is necessary in all embodiments.
The invention could be applied to a smartphone, a smart card, laptop or desktop computer, or any other electronic device with a fingerprint sensor, such as an IoT device.
FPC has had several patent filings related to fingerprint sensors on personal electronic devices published recently, including one for controlling a large fingerprint sensing area. The company also recently released the second generation of its T-Shape sensor for contactless biometric payment cards.