Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM61962
Volume:
10 |
4 |
April 2026
Pages:
1-4
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Department of Computer Engineering
Matoshri College of Engineering and Research Centre, Nashik, Maharashtra, India
Department of Computer Engineering
Matoshri College of Engineering and Research Centre, Nashik, Maharashtra, India
Department of Computer Engineering
Matoshri College of Engineering and Research Centre, Nashik, Maharashtra, India
Department of Computer Engineering
Matoshri College of Engineering and Research Centre, Nashik, Maharashtra, India
Department of Computer Engineering
Matoshri College of Engineering and Research Centre, Nashik, Maharashtra, India
This paper presents an AI-based system for generating human facial images from textual descriptions to assist in forensic investigations. Traditional suspect identification relies on manual sketching techniques, which are time consuming and dependent on artistic skills. The proposed system integrates Natural Language Processing (NLP) with Generative Adversarial Networks (GANs) to convert descriptive text into realistic facial images. Additionally, the generated images are matched against a centralized database using cloud-based facial recognition services to identify potential suspects. The system supports multilingual input and enables faster and more accurate identification compared to conventional methods. Experimental analysis highlights the effectiveness of combining text-based generation with automated recognition. This approach improves investigation efficiency and demonstrates the potential of AI-driven solutions in law enforcement and security applications.
KEYWORDS
This work demonstrates a novel approach to generating realistic human faces from fine-grained textual descriptions using a text-conditional GAN. By converting CelebA attributes into natural captions and applying training stabilizations, our method achieves promising results. The study also highlights the limitations of existing evaluation metrics, emphasizing the need for better text-to-face assessment methods.
[1] D. M. A. Ayanthi, & S. Munasinghe. (2024). Text-to-Face Generation with StyleGAN2. International Journal of Artificial Intelligence Research, 12(3), 145–152. [2] Z. Yauri-Lozano, C. Castillo-Cara, O. Orozco-Barbosa, & A. García-Castro. (2024). GANs for Text-to-Face Synthesis: Spanish NLP Encoder Analysis. IEEE Access, 32(7), 24891–24905. [3] Hengfei Wang, Zhongqun Zhang, Yihua Cheng, & Hyung Jin Chang. (2024). TextGaze: Gaze-Controllable Face Generation with Natural Language. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, 3345–3354. [4] S. Maurizio La Cava, G. Orrù, M. Drahansky, G. L. Marcialis, & F. Roli. (2023). 3D Face Reconstruction: The Road to Forensics. Forensic Science International: Digital Investigation, 45(1), 101–110. [5] Mahanthi Sriramulu, et al. (2025). Deep Learning-Based Text-to-Image Synthesis for Criminal Face Generation. Journal of Forensic Artificial Intelligence, 4(2), 77–85.
Publication Date
4/30/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
1
Total views since publication
PDF Downloads
3
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...