We propose an alternative to obtain superior 3D representations from 2D pre-trained models via I mage-to- P oint Masked Autoencoders, named as I2P-MAE. By self-supervised pre-training, we leverage the ...

Transforming your design vision into reality can be challenging, especially when precision and detail are key. That’s where 3D scanning steps in, revolutionizing how interior design and architecture ...

Abstract We present FaceVerse, a fine-grained 3D Neural Face Model, which is built from hybrid East Asian face datasets containing 60K fused RGB-D images and 2K high-fidelity 3D head scan models. A ...