Facial Recognition for Biometric Authentication: A Literature Survey

Abstract

Within this research paper, the imperative to enhance security in biometric authentication systems is treated by analyzing zero-bit watermarking to extract distinctive identifiers from biometric images. The general goal is to hunt and study existing research on zero-bit watermarking techniques specifically created for application within biometrics and how they can generatestrongandsecuredistinctiveIDswithoutcompromisingtheauthenticityoftheoriginalbiometricdata.Thevalueaddition ofthisresearchpaperisthatitoffersasystematicreviewofexistingstateofzero-bitwatermarkingmethodsforgeneratingunique IDsforbiometricimages.Itconsolidatesprominentmethodologies,discussionsandresultswithinthenewresearcharea,reflecting its importance for enhancing biometric data security and privacy.

Introduction

???? Overview

Biometric systems (using fingerprints, iris scans, facial features) offer high security but are vulnerable to misuse and tampering. Zero-bit watermarking is emerging as a non-invasive, secure method for embedding unique IDs into biometric images without altering their visual quality.

???? Objectives

1. Highlight the importance of protecting biometric data using robust security practices.

2. Review zero-bit watermarking as a technique to generate unique identifiers from biometric data.

3. Analyze embedding/extraction methods like DWT (Discrete Wavelet Transform) and SVD (Singular Value Decomposition).

4. Demonstrate how zero-bit watermarking ensures data integrity, authenticity, and privacy.

5. Identify challenges, such as robustness, imperceptibility, and computational complexity, and suggest future research paths.

???? Literature Insights

• Traditional watermarking techniques embed visible/invisible data into media but may degrade image quality.

• Zero-bit watermarking differs by not embedding actual data, but instead generating a unique ID from the biometric features.

• Recent trends include:

  • Multimodal biometrics (e.g., face + iris) for accuracy.

  • Hybrid transform-domain techniques like DWT-SVD, DWT-DCT for enhanced robustness.

  • AI and fuzzy logic for adaptive, intelligent watermarking.

• Key challenges: maintaining recognition accuracy, standardizing testing, and enabling real-time use.

???? Methodology

Watermarked images were tested against:

• Noise (Gaussian, Salt & Pepper)

• Compression (e.g., JPEG)

• Geometric manipulation (cropping, rotation)

• Filtering and brightness changes

Evaluation metrics included:

• PSNR (Peak Signal-to-Noise Ratio) – measures image fidelity.

• SSIM (Structural Similarity Index) – assesses perceptual similarity.

• BER (Bit Error Rate) – measures data integrity.

• Normalized Correlation & Hamming Distance – quantify watermark similarity and biometric feature matching.

???? Results & Discussion

? Advantages of Zero-Bit Watermarking:

• No image quality degradation → theoretically infinite PSNR, SSIM = 1.

• No embedded payload → avoids distortions affecting biometric recognition.

• Robustness depends on feature selection (needs to withstand attacks).

• Security is enhanced through unique encrypted IDs that are hard to reverse-engineer or tamper with.

• Outperforms traditional methods by keeping the original biometric data untouched.

???? Comparison with Other Techniques:

• Traditional watermarking changes pixel values → risks distortion.

• Zero-bit watermarking maintains original data integrity → ideal for biometric systems where recognition precision is critical.

???? Future Directions

1. Identify stable, attack-resistant features for ID generation.

2. Use advanced encryption and secure ID storage systems.

3. Test against adversarial attacks and real-world manipulations.

4. Incorporate machine learning for smarter feature extraction and verification.

5. Improve real-time performance and standardized benchmarks

Conclusion

AlthoughDWTandDCTfightwellagainstgeneralimagemanipulations,SVDfightswithstabilitythroughembeddinginintrinsic properties. Hybrid approaches merge the advantages of various techniques for enhanced security. In spite of this, there are still problemswithbalancingrobustnessandimperceptibility,aswellascomputationalcomplexityofsophisticatedtechniques,ruling outreal-timeusage.Zero-BitWatermarking,arecentlyproposedtechnique,aimstosteerclearofthesetrade-offsbyobtaininga biometric watermark from biometric traits without altering the original data. Despite progress, research gaps remain, such as enhancing adversarial robustness, enhancing real-time applicability, evaluating impact on recognition performance, creating standardmetrics,andexploringAI-assistedadvancements.Subsequentresearchshouldfocusontheguaranteeofsecureidentifier storage and boosting security to ward off prospective attacks to tap into the maximum capabilityof traditional and Zero-Bit watermarking to secure biometric data.

References

[1] NadaFadhilMohammed,MajidJabbarJawad,andSuhadAAli,Biometric-basedmedicalwatermarkingsystemfor verifying privacy and source authentication, Kuwait Journal of Science, Vol. [2] Chander Kant and SheetalChaudhary, \"A watermarking- based approach for protection of templates in multimodal biometrics system,\" Procedia Computer Science, Vol. 167, pp. 932–941, 2020.3 [3] Sinan Q. Salih, Ravi Sekhar, Jamal FadhilTawfeq, Amer Ibrahim, Pritesh Shah, and Ahmed Dheyaa, \"Integrated Digital Signature Based Watermarking Technology for Securing Online Electronic Documents,\" Fusion: Practice and Applications, Vol. 14, No. 01, pp. 120-128, 2024.4 [4] TaskeenTaj and ManashSarkar, \"A Survey on Embedding Iris Biometric Watermarking for User Authentication,\" [5] Cloud Computing and Data Science, Vol. 4, Iss. 2, pp. 203-211, 2023.5 [6] Om Prakash Singh, A. K. Singh, GautamSrivastava, and Neeraj Kumar, \"Image Watermarking using soft computing techniques: A comprehensive review survey,\" Multimedia Tools and Applications, DOI: 10.1007/s11042-020- 09606-x, August 2021.6 [7] Mohammed A. M. Abdullah, Satnam S. Dlay, Wai L. Woo, and Jonathon A. Chambers, \"A Framework for Iris Biometrics Protection: A Marriage between Watermarking and Visual Cryptography,\" IEEE Access, [Information inferred from the text suggests it\'s an IEEE publication, though specific volume/issue/pages are not directly stated. If this is critical, please provide more precise citation details for this source].7 [8] S. D. Shinde and P. S. Malge, \"Efficient reversible watermarking technique for secure data embedding protected templates of biometric authentication system,” international journal of innovations in engineering research and technology [IJIERT], Vol. 5, Iss. 5, May-2018.8 [9] Kumar, et al., [Information inferred from the text suggests a conference or journal paper on Rubik\'s cube scrambling for image encryption, though full details are not provided in this excerpt].9 [10] G. Balamurugan, et al., [Information inferred from the text suggests a paper on iris verification using IrisMatch-CNN, though full details are not provided in this excerpt].10 [11] M. A. M. Abdullah et al., [Information inferred from the text suggests a paper on multimodal iris recognition using deep learning, though full details are not provided in this excerpt].11 [12] Lydia Elizabeth et al., [Information inferred from the text suggests a paper reviewing watermarking system frameworks and the effectiveness of DWT, though full details are not provided in this excerpt].12 [13] M. K. Dutta, et al., [Information inferred from the text suggests a paper on distinguishing forged and original printed documents, though full details are not provided in this excerpt].

Copyright

Copyright © 2025 Prof. Poornima M, Noor Saniya, Preethi J, Ruthu S, Yogitha C. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.