Meta knocks down the cost of digital people, ordinary people can generate 3D digital people by taking a picture with their cell phones

Meta introduces new technology Avat3r, which requires only 4 facial photos taken by cell phone to generate a movable 3D digital human model in 5 minutes; adopts Dynamic 3D Gaussian Reconstruction Model + ViT Visual Transformer Architecture, combined with DUSt3R and Sapiens technology, real-time rendering can be realized by a single RTX3090 graphics card; supports single drawing/painting/sculpture Cross-dimensional modeling, micro-expression capture through FLAME facial encoding and cross attention layer, significantly reducing the threshold of digital human production.