Synthetic Data Generation for Emotional Depth Faces: Optimizing Conditional DCGANs via Genetic Algorithms in the Latent Space and Stabilizing Training with Knowledge Distillation

TL;DR

This paper introduces a novel synthetic depth face generation framework that combines optimized GANs with knowledge distillation and genetic algorithms to enhance diversity, quality, and emotional depth in facial datasets.

Contribution

It presents a new method integrating Knowledge Distillation and Genetic Algorithms to improve GAN training stability, diversity, and quality for emotional depth face synthesis.

Findings

Outperforms GAN, VAE, GMM, and KDE in diversity and quality metrics.

Achieves 94% and 96% accuracy in emotion classification.

Shows consistent improvements in FID, IS, SSIM, and PSNR metrics.

Abstract

Affective computing faces a major challenge: the lack of high-quality, diverse depth facial datasets for recognizing subtle emotional expressions. We propose a framework for synthetic depth face generation using an optimized GAN with Knowledge Distillation (EMA teacher models) to stabilize training, improve quality, and prevent mode collapse. We also apply Genetic Algorithms to evolve GAN latent vectors based on image statistics, boosting diversity and visual quality for target emotions. The approach outperforms GAN, VAE, GMM, and KDE in both diversity and quality. For classification, we extract and concatenate LBP, HOG, Sobel edge, and intensity histogram features, achieving 94% and 96% accuracy with XGBoost. Evaluation using FID, IS, SSIM, and PSNR shows consistent improvement over state-of-the-art methods.