Explore our debut article on Medium, The Power of Synthetic Data: Using Game Engines to Improve AI,co-authored by AI Research Scientist Shaghayegh (Sherry) Taheri, and AI Scientific Director Vahe Vardanyan.
Computer vision in its broadest sense allows us to perceive, analyze, recognize and process components in visual media. Recently computer vision is done mostly using AI (deep learning) techniques. At the CDRIN, it's a natural fit for our graphics and AI teams to work together and leverage synthetic data generated in game engines to train computer vision models. We use those models to help our partners in many different industries to do research and make better products.
With our expertise in computer vision and AI, we can help businesses create new tools and products and drive their success forward.
Rendering is the application of visual effects to a 3D scene and is often based on effects of physics (e.g.: reflection, refraction, radiosity, etc.). The use of shaders is often necessary to obtain real-time performance (e.g. rendering in video games).
The study of animation is the primary expertise of the CDRIN. Initially done using animation data acquisition technologies (MOtion CAPture), the team now exploits video and machine learning to understand, track, classify or compare movements or animations.
At the CDRIN, modeling is about using the skills of a programmer to reconstruct virtual objects (e.g. characters, environments, etc.) from lines of code. Modeling means starting from 2D media such as photographs, images or videos to give them volume (3D).
We have a strong expertise in automating and improving production pipelines in VFX, video games and animation. We automate exchanges between software, add features to existing software such as Unity, and conduct research to add pipeline features, including rendering and animations.
From ideation to commercialization
Technology Monitoring & Foresight
Proposition of practical applications on analytical bases.
Studies and recommendations on the basic properties of a technology applicable to a problem.
Proof of concept
The process of proving that an idea will work once it is implemented.
Analytical or laboratory studies demonstrate feasibility with individual components that are not yet integrated or representative of the final product, then the individual elements are combined to validate that they work together.
A prototype has the full technical and operating characteristics of a solution. Projects at this stage integrate, adapt and validate a complete solution in the partner's environment until the model or prototype represents a desired configuration that can be integrated into this production environment.
Production is the stage where the solution is really implemented in its final form and function. At this stage, projects consist of development and implementation for end users.
Technology Transfert & Specialized Training
You have a research result in hand, ready for the commercialization phase? To implement this solution within your company or to enhance it in another sector of activity, it is sometimes necessary to invest in skill development. In addition to ensuring that this solution is well integrated into your offer, we can also help your team prepare to exploit it.
Sign up for our newsletter to get the latest digital intelligence news, our most recent imaging case studies and essays from our experts. (In French)