Synthesis AI, a San Francisco-based startup, has unveiled a revolutionary new way of turning text prompts into lifelike 3D digital humans. This breakthrough text-to-3D technology, showcased on their online platform Synthesis Labs, uses cutting-edge generative artificial intelligence (AI) and visual effects pipelines to create high-resolution human models fit for gaming, virtual reality, film, and industrial or manufacturing simulations.
Synthesis AI has created a program that is able to create a 3D digital human from text prompts such as age, gender, ethnicity, hairstyle, and clothing. Additionally, users can edit the 3D model with text prompts or sliders capable of manipulating facial expressions and lighting. This technology offers the world an easier, quicker way to generate digital human data for machine learning (ML) training.
Synthesis AI has also developed Synthesis Humans and Synthesis Scenarios, two cutting-edge synthetic data products available in the market. With generative AI and cinematic VFX pipelines, they produce perfectly labeled synthetic data to train ML models. But it wasn’t until the launch of Synthesis Labs that Synthesis AI’s text-to-3D digital human synthesis was demonstrated in high-resolution, cinematic quality.
By introducing prompt-based input and editing, Synthesis Labs makes the 3D generative AI capabilities more accessible to developers of all skill levels. This new feature is expected to reduce the costs of 3D applications in many industries and expedite the development process.
Synthesis AI is working towards a mission to provide perfectly labeled synthetic data to support advanced AI applications. Their proprietary library of digital humans created from this text-to-3D technology is used to train models in areas such as face identification, driver monitoring, avatars, and more.
Yashar Behzadi, CEO and founder of Synthesis AI, spoke of the company’s achievements. He pointed out that the text-to-3D technology combined with the industry-standard library of digital humans is a milestone in creating synthetic data for real-world simulations. Furthermore, creating a high-quality human character costs much less than recruiting and obtaining waivers, and can be done much faster – saving both time and tens of thousands of dollars.
No code is needed to be able to generate 3D digital humans, allowing users of all experience levels to benefit from Synthesis Labs. The possibilities are endless with this new technology – from game and animation development, to VR and AR, to computer vision with synthetic data.
At its core, Synthesis AI is an advanced display tech company that provides infrastructure and services in order to help developers tightly integrate their designs and autonomous systems. Their mission is to create this bridge between the physical and digital world, enabling the design of powerful applications that simulate reality.
Behzadi also mentioned that Synthesis AI is creating an open, accessible platform that promotes collaboration and partnerships. Focus is also given to continual refinement and improvement based on user feedback.
Synthesis AI is bustling with excitement as it continues to innovate and lead the way forward in synthetic data, or perfectly labeled data, to train computer vision models. Indeed, the launch of Synthesis Labs is an immense step forward to achieve an immersive digital-physical world.
