Beyond First-Order: A Multi-Scale Approach to Finger Knuckle Print Biometrics
Chengrui Gao, Ziyuan Yang, Andrew Beng Jin Teoh, Min Zhu
TL;DR
This work tackles the limitations of first-order texture descriptors in finger knuckle print (FKP) biometrics by introducing DOTCNet, a multi-scale network that jointly models first- and second-order textures through Dual-order Texture Competitive Modules (DTCMs) with learnable Gabor filters (LGFs). Each DTCM operates on a distinct scale to extract detailed texture features via first-order LGF and structural features via second-order LGF, with Triplet Attention refining salient details and a Softmax-based competitive code suppressing noise in structure features. The architecture fuses features across three scales, producing a rich representation $F = \mathrm{FC}(\mathrm{FC}(\mathrm{Concat}(F_{ls}, F_{ms}, F_{ts})))$, and is validated on the PolyU-FKP dataset where it achieves state-of-the-art performance with an overall EER of $2.186\%$. Ablation studies demonstrate the benefits of multi-scale branches and the dual-order feature design, highlighting that TAM improves first-order details while CM effectively discriminates second-order structures, indicating robust gains for practical FKP recognition systems.
Abstract
Recently, finger knuckle prints (FKPs) have gained attention due to their rich textural patterns, positioning them as a promising biometric for identity recognition. Prior FKP recognition methods predominantly leverage first-order feature descriptors, which capture intricate texture details but fail to account for structural information. Emerging research, however, indicates that second-order textures, which describe the curves and arcs of the textures, encompass this overlooked structural information. This paper introduces a novel FKP recognition approach, the Dual-Order Texture Competition Network (DOTCNet), designed to capture texture information in FKP images comprehensively. DOTCNet incorporates three dual-order texture competitive modules (DTCMs), each targeting textures at different scales. Each DTCM employs a learnable texture descriptor, specifically a learnable Gabor filter (LGF), to extract texture features. By leveraging LGFs, the network extracts first and second order textures to describe fine textures and structural features thoroughly. Furthermore, an attention mechanism enhances relevant features in the first-order features, thereby highlighting significant texture details. For second-order features, a competitive mechanism emphasizes structural information while reducing noise from higher-order features. Extensive experimental results reveal that DOTCNet significantly outperforms several standard algorithms on the publicly available PolyU-FKP dataset.