A Review on Talk-able Facial Construction
Rishit Dubey, Nishant Doshi · 2023
A realistic video and audio portrayal is presented as a portrait image. While most current methods for creating photo-realistic video focus on breaking down information in a single image or understanding the relationship between frames, they do not effectively address the connection between audio and video information during synthesis. The use of Asymmetric Mutual Information Estimation (AMIE) and a Dynamic Attention (DA) block during training can enhance the lip synchronization in the generated video. The goal of talking face generation is to produce a seamless video of a person's face that accurately matches the movement of their lips to a given speech clip and facial image. In this paper, we have given the analysis of available approaches in talkable facial construction.