1 paper
Chubin Zhang, Hongliang Song, Yi Wei +3
In this work, we introduce the Geometry-Aware Large Reconstruction Model (GeoLRM), an approach which can predict high-quality assets with 512k Gaussians and 21 input images in only…