Impact Factor (2025): 6.9
DOI Prefix: 10.47001/IRJIET
Vol 9 No 6 (2025): Volume 9, Issue 6, June 2025 | Pages: 128-132
International Research Journal of Innovations in Engineering and Technology
OPEN ACCESS | Research Article | Published Date: 18-06-2025
The field of 3D animation is undergoing a transformative shift through deep learning. This paper presents a robust system that automates 3D facial animation from a static image using Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs). With real-time deployment on cloud platforms, the approach bypasses the need for manual animation and motion-capture tools. Our system accepts a static face image and a driving video, mapping expressions and movements with high fidelity. Performance metrics such as SSIM (0.87), FID (23.4), and 23 FPS illustrate its capability. The proposed framework is adaptable to human, cartoon, and avatar faces, demonstrating a scalable path toward democratized 3D animation.
3D Animation, Deep Learning, GAN, VAE, Real-Time Image Synthesis, Facial Animation
Sanskruti P. Upadhyay, Noor Siddiqui, Pranay Vitekar, Mahek Pathan, Pushpa Tandekar. (2025). 3D Animation Generation & Image Enhancement Using Deep Learning. International Research Journal of Innovations in Engineering and Technology - IRJIET, 9(6), 128-132. Article DOI https://doi.org/10.47001/IRJIET/2025.906015
This work is licensed under Creative common Attribution Non Commercial 4.0 Internation Licence