Snap Inc.

Adobe, Inc.

Pinscreen, Inc.

Kyle Olszewski

Sergey Tulyakov

Menglei Chai

Jian Ren

Hsin-Ying Lee

Zeng Huang

Aliaksandr Siarohin

Huan Wang

Jose Ignacio Echevarria Vallespi

Zhili Chen

Duygu Ceylan Aksit

Oliver Woodford

Hao Li

Ligong Han

Ivan Skorokhodov

Willi Menapace

Francesco Barbieri

Zhengfei Kuang

Joseph J Lim

Panagiotis Achlioptas

Evangelos Ntavelis

Zezhou Cheng

Shervin Minaee

Systems and methods for generating static and articulated 3D assets are provided that include a 3D autodecoder at their core. The 3D autodecoder framework embeds properties learned from the target dataset in the latent space, which can then be decoded into a volumetric representation for rendering view-consistent appearance and geometry. The appropriate intermediate volumetric latent space is then identified and robust normalization and de-normalization operations are implemented to learn a 3D diffusion from 2D images or monocular videos of rigid or articulated objects. The methods are flexible enough to use either existing camera supervision or no camera information at all—instead efficiently learning the camera information during training. The generated results are shown to outperform state-of-the-art alternatives on various benchmark datasets and metrics, including multi-view image datasets of synthetic objects, real in-the-wild videos of moving people, and a large-scale, real video dataset of static objects.

Autodecoding latent 3D diffusion models

Methods and systems are disclosed for performing operations for generating a 3D model of a scene. The operations include: receiving a set of two-dimensional (2D) images representing a first view of a real-world environment; applying a machine learning model comprising a neural light field network to the set of 2D images to predict pixel values of a target image representing a second view of the real-world environment, the machine learning model being trained to map a ray origin and direction directly to a given pixel value; and generating a three-dimensional (3D) model of the real-world environment based on the set of 2D images and the predicted target image.

3D modeling based on neural light field

Unsupervised volumetric 3D animation (UVA) of non-rigid deformable objects without annotations learns the 3D structure and dynamics of objects solely from single-view red/green/blue (RGB) videos and decomposes the single-view RGB videos into semantically meaningful parts that can be tracked and animated. Using a 3D autodecoder framework, paired with a keypoint estimator via a differentiable perspective-n-point (PnP) algorithm, the UVA model learns the underlying object 3D geometry and parts decomposition in an entirely unsupervised manner from still or video images. This allows the UVA model to perform 3D segmentation, 3D keypoint estimation, novel view synthesis, and animation. The UVA model can obtain animatable 3D objects from a single or a few images. The UVA method also features a space in which all objects are represented in their canonical, animation-ready form. Applications include the creation of lenses from images or videos for social media applications.

Unsupervised volumetric animation

A system to enable 3D hair reconstruction and rendering from a single reference image which performs a multi-stage process that utilizes both a 3D implicit representation and a 2D parametric embedding space.

Single image three-dimensional hair reconstruction

A multimodal video generation framework (MMVID) that benefits from text and images provided jointly or separately as input. Quantized representations of videos are utilized with a bidirectional transformer with multiple modalities as inputs to predict a discrete video representation. A new video token trained with self-learning and an improved mask-prediction algorithm for sampling video tokens is used to improve video quality and consistency. Text augmentation is utilized to improve the robustness of the textual representation and diversity of generated videos. The framework incorporates various visual modalities, such as segmentation masks, drawings, and partially occluded images. In addition, the MMVID extracts visual information as suggested by a textual prompt.

Video synthesis via multimodal conditioning

Three-dimensional object representation and re-rendering systems and methods for producing a 3D representation of an object from 2D images including the object that enables object-centric rendering. A modular approach is used that optimizes a Neural Radiance Field (NeRF) model to estimate object geometry and refine camera parameters and, then, infer surface material properties and per-image lighting conditions that fit the 2D images.

Object-centric neural decomposition for image re-rendering

A system and method are described for generating 3D garments from two-dimensional (2D) scribble images drawn by users. The system includes a conditional 2D generator, a conditional 3D generator, and two intermediate media including dimension-coupling color-density pairs and flat point clouds that bridge the gap between dimensions. Given a scribble image, the 2D generator synthesizes dimension-coupling color-density pairs including the RGB projection and density map from the front and rear views of the scribble image. A density-aware sampling algorithm converts the 2D dimension-coupling color-density pairs into a 3D flat point cloud representation, where the depth information is ignored. The 3D generator predicts the depth information from the flat point cloud. Dynamic variations per garment due to deformations resulting from a wearer's pose as well as irregular wrinkles and folds may be bypassed by taking advantage of 2D generative models to bridge the dimension gap in a non-parametric way.

3D garment generation from 2D scribble images

Systems, computer readable media, and methods herein describe an editing system where a three-dimensional (3D) object can be edited by editing a 2D sketch or 2D RGB views of the 3D object. The editing system uses multi-modal (MM) variational auto-decoders (VADs)(MM-VADs) that are trained with a shared latent space that enables editing 3D objects by editing 2D sketches of the 3D objects. The system determines a latent code that corresponds to an edited or sketched 2D sketch. The latent code is then used to generate a 3D object using the MM-VADs with the latent code as input. The latent space is divided into a latent space for shapes and a latent space for colors. The MM-VADs are trained with variational auto-encoders (VAE) and a ground truth.

Cross-modal shape and color manipulation

Systems and methods herein describe a motion retargeting system. The motion retargeting system accesses a plurality of two-dimensional images comprising a person performing a plurality of body poses, extracts a plurality of implicit volumetric representations from the plurality of body poses, generates a three-dimensional warping field, the three-dimensional warping field configured to warp the plurality of implicit volumetric representations from a canonical pose to a target pose, and based on the three-dimensional warping field, generates a two-dimensional image of an artificial person performing the target pose.

Flow-guided motion retargeting

Certain embodiments involve synthesizing image content depicting facial hair or other hair features based on orientation data obtained using guidance inputs or other user-provided guidance data. For instance, a graphic manipulation application accesses guidance data identifying a desired hair feature and an appearance exemplar having image data with color information for the desired hair feature. The graphic manipulation application transforms the guidance data into an input orientation map. The graphic manipulation application matches the input orientation map to an exemplar orientation map having a higher resolution than the input orientation map. The graphic manipulation application generates the desired hair feature by applying the color information from the appearance exemplar to the exemplar orientation map. The graphic manipulation application outputs the desired hair feature at a presentation device.

Synthesizing hair features in image content based on orientation data from user guidance

There is disclosed a system and method for training a set of expression and neutral convolutional neural networks using a single performance mapped to a set of known phonemes and visemes in the form predetermined sentences and facial expressions. Then, subsequent training of the convolutional neural networks can occur using temporal data derived from audio data within the original performance mapped to a set of professionally-created three dimensional animations. Thereafter, with sufficient training, the expression and neutral convolutional neural networks can generate facial animations from facial image data in real-time without individual specific training.

Growing community of inventors

Los Angeles, CA, United States of America

Kyle Olszewski