Synthesis AI has released new products for human-centric computer vision models | Tech Rasta


Synthesis AI, a leader in synthetic data technologies, announced the release of two new products for the development of advanced human-centric computer vision models. The two products, called Synthesis Humans and Synthesis Scenarios, are powered by generative AI and cinematic computer graphics technologies and enable the simulation of complex multi-human data.

According to the company, the new products are an extension of the company’s existing data product platform. They are designed to make the creation and implementation of synthetic data easier and more accessible.

Yashar Behzadi is the CEO and founder of Synthesis AI.

“Synthesis Humans and Synthesis Scenarios are a natural evolution of our synthetic data roadmap,” Behzadi said. “Synthetic data powered by generative AI is now recognized as a more efficient paradigm for building computer vision AI. Our new products enable the development of more sophisticated multi-human AI models required by emerging applications.”

Synthetic humans and synthetic visual abilities

By using Synthesis Humans, machine learning developers can build advanced production-level models. The product provides more than 100,000 unique identifiers and gives learners the ability to edit various attributes such as emotion, body type and motion. The user interface allows them to seamlessly create labeled data, and teams that prefer programmatic access and control can use the comprehensive API.

The other major product, Synthesis Scenarios, is the first to allow fine-grained control of complex, multi-human simulations in a wide range of environments for machine learning model development. Various companies involved in Augmented Reality (AR), Virtual Reality (VR), Autonomy and more rely on the platform to create efficient and robust models.

Supports a wide range of applications

The new products mean Synthesis AI now supports machine learning model development for a wide range of existing and emerging applications such as ID verification, AR/VR, Metaverse, virtual try-on, AI fitness, driver and passenger monitoring, teleconferencing, pedestrian detection, etc. , VFX, Security and more.

For example, a company involved with ID verification can use Synthesis Humans to access a variety of privacy-compliant data, which can help reduce privacy and regulatory constraints in obtaining facial data. When it comes to AR, VR, and Metaverse, Synthesis Humans can be used to label 3D data across a wide range of populations, while Synthesis Scenes supports the development of multi-person tracking and interaction models.

These are just a few of the many potential applications of both products.

Florian Leibert of 468 Capital had this to say about the company and its offerings:

‚ÄúSynthesis AI continues to innovate in synthetic data. The company’s platform is the first to bring generative AI together with computer graphics pipelines to help their customers create photorealistic, differentiated and accurately labeled training data. The new products represent a significant advance, enabling computer vision teams to more cost-effectively build ML models across more applications and industries.

Both new products will be available from this month (November 2022).


Source link