Snap Inc.

Hsin-Ying Lee

Sergey Tulyakov

Menglei Chai

Jian Ren

Kyle Olszewski

Aliaksandr Siarohin

Ivan Skorokhodov

Gurunandan Krishnan Gorumkonda

Zeng Huang

Willi Menapace

Jie Xu

Yinghao Xu

Aleksei Stoliar

Sergei Korolev

Ligong Han

Zezhou Cheng

Shervin Minaee

Francesco Barbieri

Panagiotis Achlioptas

Rameen Abdal

Peihao Zhu

Chieh Lin

Yuwei Fang

Songfang Han

Tsai-Shien Chen

Automatic captioning pipelines and methods for automatically annotating video data with subtitles, which can be obtained using automatic speech recognition (ASR). An automatic captioning pipeline with inputs of multimodal data scales up the dataset of high-quality video-caption pairs. The automatic captioning pipeline generates video-caption pairs by establishing and using a large video-language dataset along with an automatic captioning approach leveraging multimodal inputs, such as textual video description, subtitles, and individual video frames.

Captioning videos with multiple cross-modality teachers

An artificial intelligence (AI) network or neural network is trained to generate three-dimensional (3D) models or shapes with color from two-dimensional (2D) input images and input text describing the 3D model with color. Example methods include converting a first three-dimensional (3D) model from a first representation to a second representation, the second representation including color information for the 3D model and inputting the second representation into an encoder to generate a third representation having a lower dimension than the second representation. The method further includes inputting the third representation into a decoder to generate a fourth representation having a same dimension as the second representation and generating a second 3D model from the fourth representation. The method further includes determining losses between the first 3D model and the second 3D model and updating weights of the encoder and the decoder based on the losses.

Generating 3D models with texture

A 3D-aware generative model for high-quality and controllable scene synthesis uses an abstract object-level representation (i.e., 3D bounding boxes without semantic annotation) as the scene layout prior, which is simple to obtain, general to describe various scene contents, and yet informative to disentangle objects and background. An overall layout for the scene is identified and then each object is located in the layout to facilitate the scene composition process. The object-level representation serves as an intuitive user control for scene editing. Based on such a prior, the system spatially disentangles the whole scene into object-centric generative radiance fields by learning on only 2D images with global-local discrimination. Once the model is trained, users can generate and edit a scene by explicitly controlling the camera and the layout of objects' bounding boxes.

Spatially disentangled generative radiance fields for controllable 3D-aware scene synthesis

An environment synthesis framework generates virtual environments from a synthesized two-dimensional (2D) satellite map of a geographic area, a three-dimensional (3D) voxel environment, and a voxel-based neural rendering framework. In an example implementation, the synthesized 2D satellite map is generated by a map synthesis generative adversarial network (GAN) which is trained using sample city datasets. The multi-stage framework lifts the 2D map into a set of 3D octrees, generates an octree-based 3D voxel environment, and then converts it into a texturized 3D virtual environment using a neural rendering GAN and a set of pseudo ground truth images. The resulting 3D virtual environment is texturized, lifelike, editable, traversable in virtual reality (VR) and augmented reality (AR) experiences, and very large in scale.

Infinite-scale city synthesis

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 three-dimensional (3D) scene is generated from non-aligned generic camera priors by producing a tri-plane representation for an input scene received in random latent code, obtaining a camera posterior including posterior parameters representing color and density data from the random latent code and from generic camera priors without alignment assumptions, and volumetrically rendering an image of the input scene from the color and density data to provide a scene having pixel colors and depth values from an arbitrary camera viewpoint. A depth adaptor processes depth values to generate an adapted depth map that bridges domains of rendered and estimated depth maps for the image of the input scene. The adapted depth map, color data, and scene geometry information from an external dataset are provided to a discriminator for selection of a 3D representation of the input scene.

3D generation of diverse categories and scenes

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

Domain adaptation frameworks for producing a 3D avatar generative adversarial network (GAN) capable of generating an avatar based on a single photographic image. The 3D avatar GAN is produced by training a target domain using an artistic dataset. Each artistic dataset includes a plurality of source images, each associated with a style type, such as caricature, cartoon, and comic. The domain adaptation framework in some implementations starts with a source domain that has been trained according to a 3D GAN and a target domain trained with a 2D GAN. The framework fine-tunes the 2D GAN by training it with the artistic datasets. The resulting 3D avatar GAN generates a 3D artistic avatar and an editing module for performing semantic and geometric edits.

Avatar generation according to artistic styles

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

Example methods for generating an animated character in dance poses to music may include generating, by at least one processor, a music input signal based on an acoustic signal associated with the music, and receiving, by the at least one processor, a model output signal from an encoding neural network. A current generated pose data is generated using a decoding neural network, the current generated pose data being based on previous generated pose data of a previous generated pose, the music input signal, and the model output signal. An animated character is generated based on a current generated pose data; and the animated character caused to be displayed by a display device.

Music reactive animation of human characters

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.

Growing community of inventors

San Jose, CA, United States of America

Hsin-Ying Lee