Machine Learning Approaches for Gender Identification Using Fingerprint Biometrics: Role of Ridge Flow, Minutiae, and Multi-resolution DWT Features | Chapter 2 | New Horizons of Science, Technology and Culture Vol. 3

Fingerprints serve as an extraordinary biological signature that encapsulates both identity and subtle gender-specific characteristics through their complex structural and spectral properties. This study introduces an advanced computational framework for gender classification by synergistically combining three distinct fingerprint feature domains: ridge geometry for macroscopic pattern analysis, minutiae distribution for microscopic feature examination, and frequency decomposition through sophisticated wavelet transformation. The methodology processes high-resolution fingerprint images through a multi-stage analytical pipeline that precisely quantifies ridge length variations (capturing minimum, maximum, and average measurements), systematically enumerates minutiae points (including ridge terminations and bifurcations), and performs multi-resolution spectral analysis using a six-level discrete wavelet transform to isolate discriminative frequency components. Validated on a carefully balanced dataset of 100 subjects (50 males and 50 females), the extracted features are intelligently organised into gender-specific clusters through an optimised stratification process, yielding an impressive overall classification accuracy of 88.28%. Notably, the right ring finger demonstrated exceptional diagnostic performance with 95.46% accuracy, a finding consistent with established embryological research on androgen-influenced ridge formation patterns. The technical sophistication of this approach lies in its ability to achieve high accuracy without relying on computationally intensive deep learning architectures, making it particularly suitable for real-world applications where efficiency is crucial. Beyond its immediate results, this research opens promising avenues for future investigation, including the incorporation of additional discriminative features such as sweat pore distribution and three-dimensional ridge curvature analysis, as well as expansion to larger, more diverse demographic datasets to enhance generalizability. By demonstrating that fingerprints contain a wealth of underutilised gender information, this work makes a significant contribution to the field of soft biometrics, with important implications for forensic science, security systems, and demographic research, while simultaneously establishing a foundation for future exploration of ancillary biometric markers embedded within fingerprint patterns. The balanced integration of robust methodology, empirical validation, and practical applicability positions this research as a valuable reference point for both academic investigation and applied biometric solutions, bridging the gap between theoretical innovation and real-world implementation in the evolving landscape of gender classification technologies.

 

Author(s) Details

Sayed Abulhasan Quadri
SECAB Institute of Engineering and Technology, Vijayapura, India.

 

Chandrakant P. Divate
SECAB Institute of Engineering and Technology, Vijayapura, India.

Tabasum Guledgudd
SECAB Institute of Engineering and Technology, Vijayapura, India.

Sayed Abdulhayan
PACE College, Mangalore, India.

 

Please see the book here:- https://doi.org/10.9734/bpi/nhstc/v3/5745