A Django-Enabled Hybrid Framework for Intelligent Face Morph Synthesis and Authentication Resilience

Abstract

Facial recognition systems are widely used for identity verification but are vulnerable to face morphing attacks, where multiple facial images are blended to form a deceptive identity that can fool recognition models. This project develops a deep learning-based approach to detect such attacks and strengthen biometric authentication systems. It lies in the domain of Artificial Intelligence and Machine Learning, focusing on Computer Vision techniques to differentiate real and morphed facial images for accurate and secure verification. The project involves creating realistic morphed face datasets and building an efficient detection model applicable to border control, ID verification, and digital authentication. Current systems fail against high-quality morphs produced using advanced tools, showing reduced accuracy under variations in lighting, age, and facial accessories. To overcome this, the proposed model combines deep learning-based feature extraction with machine learning classifiers. Morph-2 and Morph-3 datasets are generated using professional morphing tools, and image enhancement with feature fusion is applied to improve accuracy and robustness.

Introduction

The study proposes a web-based intelligent framework for generating and detecting face morphing attacks to improve the security of biometric authentication systems. With the increasing use of facial recognition for identity verification in airports, immigration, and secure facilities, face morphing has emerged as a serious threat. A morphing attack combines facial features from two individuals into a single realistic image, potentially allowing unauthorized access by deceiving both humans and automated recognition systems. The proposed system aims to accurately distinguish genuine facial images from morphed ones using image processing and machine learning techniques.

The literature review examines recent advances in face morphing attack detection. Existing approaches include deep de-morphing frameworks using StyleGAN and ArcFace, Vision Transformer (ViT) models for single-image detection, Selective Kernel Networks (SKNet) combined with Error-Level Analysis (ELA), Explainable AI (XAI) techniques using EfficientNet and Grad-CAM, interpretable deep learning methods for identity separation, and video-based morph detection (V-MAD) that analyzes facial motion across multiple frames. These studies demonstrate significant improvements in detection accuracy, robustness, and interpretability.

The proposed system integrates face morph generation, morph attack detection, image similarity analysis, and a Django-based web application into a unified platform. The morph generation module combines two facial images after preprocessing and alignment to create a realistic composite image. The detection module compares uploaded images with generated morphs using the Structural Similarity Index Measure (SSIM), which evaluates luminance, contrast, and structural similarity to determine whether an image is genuine or morphed. The system is designed to handle variations in lighting, facial expressions, age, head pose, hairstyle, and image quality.

The system architecture consists of five major layers. The Input Layer accepts two facial images and performs preprocessing such as resizing, alignment, and normalization. The Morph Synthesis Layer uses the Discrete Wavelet Transform (DWT) to combine facial features and generate a morphed image. The Feature Extraction Layer employs a Convolutional Neural Network (CNN) to extract high-level facial features, including texture and structural information. The Classification Layer uses a Support Vector Machine (SVM) classifier, supported by SSIM-based similarity analysis, to classify images as either genuine or morphed. Finally, the Web Interface Layer, developed using Python, Django, OpenCV, and PyWavelets, provides a user-friendly platform for uploading images, generating morphs, performing detection, and viewing results.

Conclusion

The Face Morphing Attack Generation and Detection System was created to show the dangers of manipulating facial images and to offer a way to spot possible morphing attacks. Users can create morphed facial photos and analyse them using image similarity approaches thanks to the project\'s effective integration of morph image generation and morph detection functionalities within a single platform. The suggested approach creates realistic morph samples by combining facial features from several photos using image processing techniques. The Structural resemblance Index Measure (SSIM) is used to compare and assess the degree of resemblance across facial photographs in order to detect manipulated images. The results show that, in a variety of testing scenarios, the system can successfully distinguish between real and altered photos. The application showed dependable performance in managing image upload, morph generation, image comparison, and result generating duties during implementation and testing. The web-based interface created using Django offers a user-friendly setting for interacting with the system and effectively carrying out morph analysis. The study draws attention to the possible security issues that face facial recognition systems when they are subjected to morphing assaults. The suggested solution advances knowledge of biometric vulnerabilities and potential defences by fusing image processing methods with similarity-based analysis. All things considered, the built system accomplishes its goals and provides a useful framework for researching face morphing assaults and their detection. The testing results show the usefulness of the suggested method in biometric security research and validate its efficacy.

References

[1] F. Peng, L.-B. Zhang, and M. Long, “FD-GAN: Face De-Morphing Generative Adversarial Network for Restoring Accomplice\'s Facial Image,” IEEE Access, vol. 7, pp. 75122–75131, 2019. [2] M. Ferrara, A. Franco, and D. Maltoni, “Face Demorphing,” IEEE Transactions on Information Forensics and Security, vol. 13, no. 4, pp. 1008–1017, Apr. 2018. [3] U. Scherhag, C. Rathgeb, J. Merkle, R. Breithaupt, and C. Busch, “Face Recognition Systems Under Morphing Attacks: A Survey,” IEEE Access, vol. 7, pp. 23012–23026, 2019. [4] A. W. Yip and P. Sinha, “Contribution of Colour to Face Recognition,” Perception, vol. 31, no. 8, pp. 995–1003, 2002. [5] U.Scherhag,C. Rathgeb,J. Merkle,and C.Busch,“Deep Face Representations for Differential Morphing Attack Detection,” IEEE Transactions on Information Forensics and Security, vol.15,pp. 3625–3639, 2020. [6] W3Schools, “Python Tutorial.” [Online]. Available: https://www.w3schools.com/python/ [7] GeeksforGeeks, “Python Programming Language.” [Online]. Available: https://www.geeksforgeeks.org/python-programming-language/Python Software Foundation, “Python Official Documentation.” [Online]. Available: https://docs.python.org/3/ [8] TutorialsPoint, “Python Tutorial.” [Online]. Available: https://www.tutorialspoint.com/python/ [9] Real Python, “Python Tutorials and Resources.” [Online]. Available: https://realpython.com/ [10] Django for Beginners, “Introduction to Django.” [Online]. Available: https://djangoforbeginners.com/introduction/ [11] Guru99, “Django Tutorial.” [Online]. Available: https://www.guru99.com/django-tutorial.html

Copyright

Copyright © 2026 Mekala Pooja, N Naveen Kumar. 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.