Efficient Face Morphing and Demorphing with Explainable AI using FSGAN
Bibliographic record
Abstract
Face morphing attacks present a significant threat to biometric security systems, often enabling imposters to gain unauthorized access by blending features from multiple identities. This study suggests a new face morphing and demorphing combined with explainable AI and identity-preserving face recognition under Face Swapping GAN (FSGAN) architecture. First, the identity embeddings are projected by facial images that are encoded with alignment through a pre-trained ArcFace model to guarantee identity representation. The FSGAN generates morphing images through the interpolation of source attributes and target features with a ratio of 70:30 to generate visually appealing hybrid identities. Demorphing, in its turn, rebuilds the original identities following a latent-space inversion method and cosine similarity to check. Explainability tools such as Grad-CAM and SHAP are added to visualize the decision-making process of the face recognition subjected to morph attack conditions to maintain transparency. Experimental results on the CelebA data demonstrate that the suggested approach presents the morphing precision of 97.56%, identity recovery accuracy of 95.88%, and morph detection score of 96.45%, which beats other current models such as StarGAN and DeepFakeGAN. Moreover, the false acceptance rate in view of morphing attacks is kept to the bare minimum of 3.12%, reflecting their high security strength. It will use a binary XGBoost classifier that leverages deep and handcrafted features to differentiate real and morphed images as well. The given overall and explainable solution is indicative to the real-time deployment of biometric security. The presented FSGAN-XAI pipeline does not only offer proper morph generation, reversal, and recognition but also offers increased stability and forensic usefulness of automated identity verification systems.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".