Ravindra Yadav

Papers

Total Citations

H-Index

About

Ravindra Yadav is a leading researcher in computer vision and generative AI, with a primary focus on audio-driven facial animation and talking face generation. His most influential work, "Stochastic Talking Face Generation Using Latent Distribution Matching" (2020), addresses the fundamental challenge of generating diverse, realistic facial movements from a single audio input—a uniquely human perceptual skill. Unlike prior deterministic approaches that produced a single output, Yadav’s method introduces stochasticity through latent distribution matching, enabling a variety of plausible talking face videos from the same speech signal. This work has garnered 3 citations and laid the groundwork for more expressive and naturalistic avatars in virtual communication, entertainment, and accessibility technologies. Yadav’s contributions push the boundaries of multimodal learning, bridging audio and visual modalities to create lifelike, dynamic facial animations. His research is particularly notable for tackling the one-to-many mapping problem in generative models, offering a more flexible and human-like alternative to traditional talking face systems. For students and researchers, Yadav’s work exemplifies how latent variable models can enhance realism and diversity in generative tasks, opening new avenues for interactive media and human-computer interaction.