Contactless Fingerprint Image Streams and Heart Rate (CFISHR): A Dataset for Biometric and Liveness Detection Research

The paper presents a novel dataset of continuous, high-quality, contactless fingerprint image streams of the right-hand thumb finger captured from 46 participants, along with synchronized heart rate measurements. The presented dataset was captured with the help of an off-the-shelf monochrome blue-light fingerprint scanner of 500 ppi with 14 fps, accompanied by a commercially available smartwatch for measuring heart rates. Each participant provided one minute of right-hand thumb fingerprint video, resulting in over 38,000 grayscale images to provide for the option of extracting photoplethysmogram signals from the fingerprint stream. This multimodal dataset is a significant contribution to biometric research, enabling advancements in dynamic fingerprint recognition, liveness detection, and physiological signal analysis. Additionally, we open-sourced MATLAB scripts for noise reduction, fingertip cropping, and heart rate computation from the stream to ensure its usability and reproducibility.