Predicting the Future Appearances of Lost Children for Information Forensics with Adaptive Discriminator-Based FLM GAN

This article proposes an adaptive discriminator-based GAN (generative adversarial network) model architecture with different scaling and augmentation policies to investigate identify the cases of lost children even after several years (as human facial morphology changes specific years). Uniform probability distribution combined random auto techniques generate future appearance children’s faces are analyzed. X-flip rotation applied periodically during pixel blitting improve pixel-level accuracy. With anisotropic scaling, images were generated by generator. Bilinear interpolation was carried out up-sampling setting padding reflection geometric transformation. The four nearest data points used estimate such at a new point interpolation. color transformation Luma flip on matrices spread log-normally for saturation. luma-flip components use brightness information each as chrominance. various modifications, StyleGan ADA architecture, implemented using NVIDIA V100 GPU. FLM method yields BRISQUE score between 10 30. uses MSE, RMSE, PSNR, SSMIM parameters compare state-of-the-art models. Using Universal Quality Index (UQI), model-generated output maintains high quality. proposed obtains ERGAS (12 k–23 k), SCC (0.001–0.005), RASE (1 k–4 SAM (0.2–0.5), VIFP (0.02–0.09) overall scores.