Fingerprint biometrics can tell a lot more about an individual than who they are, as participants in a presentation by the European Association for Biometrics (EAB) heard, although in some cases the changes that make diseases and other conditions detectable by fingerprints also make them useless for identification purposes.
The EAB Luncheon on “Disease Detection and Recognition in Fingerprint Images” was presented by a Ph.D. researcher Mona Heidari from Brno University of Technology.
Heidari explained the basic structure of the skin, how the three layers are affected by different conditions, and the various factors that can affect fingerprint image quality. Damage to the inner fat layer makes it impossible to read fingerprints, for example.
Damage can be seen in the epidermis and dermis layers in the form of structural changes or discoloration, Heidari says. A variety of conditions, some commonly found in the Third World, manifest as eczema or psoriasis which can render fingerprint ridges unreadable.
The discoloration or structural changes seen on the fingerprint can sometimes indicate the type of health problem they are facing, according to Heidari. Straight lines often, but not always, indicate cut wounds, while irregular dark spots and small hairlines are some of the possible indications of diseases or other health problems.
Heidari also explained the sources used to construct the biometric dataset used in the research.
His team applied fingerprint recognition algorithms to fingerprints after the images were classified by detection algorithms.
A classic approach with preprocessed images was used, due to the low volume of data available, was used to generate histograms, which showed promise, Heidari says. Other algorithms and approaches have been successful to widely varying degrees.
Available data is not sufficient for machine learning approaches, so Heidari experimented with synthetic data, yielding first positive results.
A researcher working on infant biometrics said Biometric update Previously, indications of highly sensitive health information could be seen at a glance in certain fingerprint images, opening up a series of ethical dilemmas and data protection issues.
The next EAB event is the WIC Midwinter meeting on “AI and Privacy – Regulatory, Technical and Scientific Perspectives”, and the next lunchtime lecture will address “Why deepfakes are not the real challenge for remote biometrics”, presented by Ann-Kathrin Freiberg of BioID.
biometric data | biometric identification | biometrics | biometric research | disease detection | eBike | European Association for Biometrics | fingerprint recognition | fingerprints