AI-Driven Forensic Face Sketch Construction and Recognition

DOI

10.2991/978-94-6463-858-5_231How to use a DOI?

Keywords

Forensic face sketch; Face Recognition; Deep learning; Feature selection; Convolution Neural Network; Generative Adversarial Network; Criminal investigations

Abstract

Facial sketching and recognition is a critical crime solving tool in identifying criminal suspects. Conventionally, sketching is done using forensic artists interpreting eyewitness accounts into a drawing, a process prone to being subjective, time-consuming, and highly reliant on artistic skill. In addition, current automated methods for Face recognition try to bring accuracies but have serious limitations. Convolution Neural Network (CNN) based recognition models tend to lose fine-grained identity specific information, resulting in poor feature matching, while Generative Adversarial Networks (GAN)-based transformations increase realism but occasionally sacrifice essential facial features required for effective recognition. These shortcomings render forensic face recognition time-consuming, unreliable, and less accurate in high-risk criminal investigations where accuracy is crucial. This paper introduces an AI-based system that trans- forms forensic sketching and recognition into a precise, efficient, and reliable tool. The initial phase enables investigators to build rich composite sketches by choosing individual facial features such as eyes, nose, lips, eyebrows, and hair / beard, so that every element positively contributes to the overall identity. To enhance consistency, an AI-facilitated recommendation engine provides compatible features, advancing realism without compromising proportionality. This minimizes human bias and guarantees that the resulting sketch better resembles the suspect’s true appearance. The process of converting a sketch into a digital image preserves major identity characteristics, providing for increased visual detail and closer resemblance to actual photographs. Lastly, recognition through deep learning guarantees that the output image is properly compared to a database of prior criminals and shows the top five matches with percentages of individual feature similarity. This process-oriented method hastens investigation, boosts time, and gives law enforcement agencies a more data-dependent, accurate and scalable way to identify suspects.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Kathyayini Pasunuri
AU  - Cheera Rohan
AU  - Gillala Vaishak Reddy
AU  - Talla Sai Sree
PY  - 2025
DA  - 2025/11/04
TI  - AI-Driven Forensic Face Sketch Construction and Recognition
BT  - Proceedings of International Conference on Computer Science and Communication Engineering (ICCSCE 2025)
PB  - Atlantis Press
SP  - 2764
EP  - 2781
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-858-5_231
DO  - 10.2991/978-94-6463-858-5_231
ID  - Pasunuri2025
ER  -