S. Kim, “MoDiTalker: Motion-Disentangled Diffusion Model for High-Fidelity Talking Head Generation”, AAAI, vol. 39, no. 4, pp. 4302–4310, Apr. 2025.