Enhancing Fingerprint Image Resolution Using Auto-Encoder and Interpolation Techniques
P P Lisha, V. K. Jayasree · International Journal of Electronics and Communication Engineering · 2024
Super-Resolution (SR) techniques play a vital role in enhancing the resolution of low-quality images, including fingerprint images, which are crucial for various applications such as biometric authentication and forensic analysis. However, achieving high-quality super-resolution of fingerprint images poses several challenges, including preserving fine details and minimizing artifacts. Existing methods often struggle to effectively enhance fingerprint images without sacrificing important features or introducing unwanted distortions. This study presents a novel approach for enhancing the resolution of low-resolution fingerprint images to high-resolution ones using an auto-encoder in combination with spline and bi-cubic interpolation methods. For effective analysis in a variety of applications, including satellite photography, surveillance, and medical imaging, highquality images are imperative. However, the transmission, storage, and processing of high-resolution images require substantial bandwidth, storage, and computational resources. Therefore, there is a need for computationally efficient and size-optimized resolution enhancement algorithms. The proposed method enhances fingerprint images from a size of 129 × 97 to 258 × 194. Utilizing spline and bi-cubic interpolation techniques ensures smoother curves and reduced artifacts during image up-sampling, preserving fidelity and detail. The model is trained on the FVC 2004 dataset and tested on both FVC 2004 and FVC 2002 datasets. Performance evaluation metrics such as SSIM, PSNR, and MSE yielded values of 35.14, 0.968, and 0.007, respectively. Furthermore, the identification accuracy of the proposed model, measured using the SIFT algorithm, achieved 100% on these datasets. The outcomes of the experiment show the effectiveness and superiority of the proposed approach in enhancing fingerprint image resolution, paving the way for improved accuracy and reliability in fingerprint recognition systems.