Automatic Fingerprint Classification Using Deep Learning Technology (DeepFKTNet)

Fingerprints are gaining in popularity, and fingerprint datasets becoming increasingly large. They often captured utilizing a variety of sensors embedded smart devices such as mobile phones personal computers. One the primary issues with recognition systems is their high processing complexity, which exacerbated when they gathered using several sensors. way to address this issue categorize fingerprints database condense search space. Deep learning effective designing robust classification methods. However, architecture CNN model laborious time-consuming task. We proposed technique for automatically determining adaptive classification; it determines number filters layers Fukunaga–Koontz transform ratio between-class scatter within-class scatter. It helps design lightweight models, efficient speed up process. The method was evaluated two public-domain benchmark FingerPass FVC2004 datasets, contain noisy, low-quality obtained live scan cross-sensor fingerprints. designed models outperform well-known pre-trained state-of-the-art techniques.