Games, Movies & Media Applications

Accurate motion capture for 3D Character actors still requires tracking markers or physical augmentation–and even then, they aren’t perfect from one session to the next. With precise landmarks in 3D space for every image, our datasets assist in capturing actors’ performances with higher accuracy; even for landmarks that would normally be occluded.

Landmark & Facial Feature Estimation
Up your mo-cap game with these datasets. Because of historically low-quality, this task has been performed with a ton of manual intervention for desired levels of realism. With a plethora of identities making diverse facial expressions in harsh environments, you can reduce error and increase quality.
Landmarks Dataset
Maximum expressions and camera angles
Essentials Dataset
Bootstrap your model with this foundation
Inclusivity Dataset
With even genders, ages, and skin tones
Programmatic API
Fully customizable data for augmentation
Crafted with facial landmarks in 3D across the spectrum of facial expressions, you’ll have the data your application needs to sense emotion.
View Dataset
The essentials dataset is an extremely well-rounded variety of identities and environments to start up and augment your existing models.
View Dataset
The inclusivity dataset showcases traditionally under-represented groups to ensure that your application is as unbiased as possible.
View Dataset
Our Face API lets you specify exactly the parameters you need, generating millions of images with unique people, environments, and camera settings.
Learn about our API
Portrait Matting
Some media use cases don’t need anything as advanced as motion capture. Instead, accurate background removal with anti-aliasing is what’s critical. To develop better models, a broader set of scenarios is needed, which the datasets below provide. Let’s rid the world of bad green-screening effects once and for all.
Essentials Dataset
Bootstrap your model with this foundation
Inclusivity Dataset
With even genders, ages, and skin tones
Harsh Lighting Dataset
Handle edge cases in bright and dark light
Programmatic API
Fully customizable data for augmentation
The essentials dataset is an extremely well-rounded variety of identities and environments to start up and augment your existing models.
View Dataset
Crafted with facial landmarks in 3D across the spectrum of facial expressions, you’ll have the data your application needs.
View Dataset
The harsh lightning dataset lets you handle your poor lighting edge-cases with glare, spotlights, night time, and backlit images.
View Dataset
Our Face API lets you specify exactly the parameters you need, generating millions of images with unique people, environments, and camera settings.
Learn about our API
Domain Adapt
As with all synthetic data, there’s a shift from our domain to the one captured by real cameras. Although there’s no universal domain adaptation approach for every use-case, we stand on the shoulders of giants to get great results.
Adaptive Batch Normalization
Adaptive Batch Normalization is a simple technique, can be easily applied to any network with batch normalization layers, and combined with all other techniques for surprisingly good results.
Revisiting Batch Normalization For Practical Domain Adaptation. Domain-Specific Batch Normalization for Unsupervised Domain Adaptation. Official Implementation of “Domain Specific Batch Normalization for Unsupervised Domain Adaptation (CVPR2019)”.
Adversarial Domain Adaptation
Adversarial domain adaptation and its modifications for particular tasks usually result in strong improvement. The downside is that it typically requires heavy pipeline modifications.
Domain-Adversarial Training of Neural Networks. Alleviating Semantic-level Shift: A Semi-supervised Domain Adaptation Method for Semantic Segmentation. A collection of implementations of adversarial domain adaptation algorithms.
Refinement
Image-2-image translation methods coupled with self-regularization loss allows dataset-level refinement. While these methods require additional pipeline to train, it is completely independent and does not require modifications of the main training pipeline.
Contrastive Learning for Unpaired Image-to-Image Translation. Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks. Learning from Simulated and Unsupervised Images through Adversarial Training. Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch).
Combined methods
For the best results all the methods above typically should be combined together.
CyCADA: Cycle-Consistent Adversarial Domain Adaptation. Code to accompany ICML 2018 paper.
Ready to Grow With You
We’re here to help you create your solution with the help of our programmatic data platform.
Scales out of the box

Our technology seamlessly scales in the cloud with our customers’ demands, from R&D phases with small amounts of data to production requirements of terabytes of data.rnrnWith everything available via an API, we integrate seamlessly with your workflows from day 1.

Learn More about our API

Learn More about our API

Machine Learning Development Support

If your team needs a little more machine learning muscle, our experts are ready to jump in. We’ll help reduce your time to market, so don’t hesitate to reach out.

Contact Us

Contact Us

Synthesis AI Raises a $17 Million Series A To Expand Its Synthetic Data Platform for Computer Vision AI

Learn More
X