A Study on Structural and Textural Feature Extraction for Contactless Fingerprint Classification

Objectives: The Recent COVID 19 Pandemic has capped the efficacy of mostly used existing touch-based fingerprint detection. Hence, main objective this study is to develop a lightweight, robust and efficient touchless identification model. To cope up with environment demands, an emphasis made on improvement in local conditions, feature modalities as well learning environment. Methods: Considering these objectives, focus two different methods for classification contactless fingerprint. In first method structural features like minutiae details are extracted followed by SSIM. second GLCM textural were using Random Forest algorithm. proposed method, performance assessment done considering data samples 1000 random users that collected from benchmark databases Hong Kong Polytechnic University 3D-fngerprint images Database Version 2.0, Touchless Fingerprint IIT Bombay, Kanpur, Jodhpur. Findings: Though, detection considered viable alternative; yet, real-time complexities non-linear patterns, dusts, non-uniform conditions illumination, contrast, orientation make it complex realization. Moreover, likelihood ridge discontinuity texture damages can majorly limit its efficacy. Novelty: model mainly focusses reducing Equal Error Rate improving accuracy extracting rather just sticking conventional feature-minutiae. Proposed outperforms when compared state art achieving 94.72%, precision 98.84%, recall 97.716%, FMeasure 0.9827 reduced EER about 0.084. key novelty approach was doesn’t require any surface 3D reconstruction, employed mathematical approaches retrieve normal information. Keywords: SSIM; GLCM; Contactless Fingerprint; Minutiae; EER; Confusion Matrix