Synthetic data for forensic analysis

Exploring whether synthetic facial data can provide a valid basis for evaluating forensic biometric methods.

Reliable forensic evidence depends on methods being empirically validated using appropriate test data. In forensic biometrics, sourcing large, ethical and legally compliant datasets.

This project explored the use of generative artificial intelligence to produce synthetic facial images that could support the validation of forensic biometric methods.

The focus was on assessing whether synthetic data can provide a meaningful and reliable basis for evaluation, and how these approaches perform in practice compared to real-world data.

The project also highlights the importance of understanding the limitations of synthetic data, and how its use may affect the interpretation of results.

This work sits at the intersection of forensic science, biometrics, and AI, examining how emerging techniques can be applied while maintaining scientific validity.