Generative adversarial network-based forensic facial reconstruction from ۳D face meshes: a comparative study of DCGAN and StyleGAN

Forensic facial reconstruction is a critical investigative technique for identifying unknown individuals from skeletal remains, facial meshes, or degraded photographs. Existing approaches manual clay reconstruction, drag-and-drop composite software, and Shape-from-Shading algorithms are time consuming, cost-prohibitive, or limited to well-defined facial inputs. This paper proposes an automated, generative AI based pipeline that reconstructs realistic facial images from ۳-D face mesh representations and ۲-D facial sketches using Deep Convolutional Generative Adversarial Networks (DCGAN) and Style-based Generative Adversarial Networks (StyleGAN). A dedicated dataset of ۵ ۰۰۰ paired mesh-to-face image samples was constructed using Google MediaPipe, covering diverse age groups, ethnicities, and genders with an ۸۰/۱۰/۱۰ train-validationtest split. DCGAN achieved a testing accuracy of ۹۴.۲% at ۳۰۰ epochs, while StyleGAN attained a Fr´echet Inception Distance (FID) of ۱۸.۳ at ۲۰۰ epochs with SSIM = ۰.۷۸ and PSNR = ۲۸.۱ dB. Comparative evaluation shows that DCGAN excels at rapid broad-pattern reconstruction, while StyleGAN provides superior feature-level perceptual fidelity. Ethical risks including misidentification, dataset consent, demographic bias, and legal implications are also addressed.

کلیدواژه ها:

Facial Reconstruction ، StyleGAN ، DCGAN ، Forensic AIF