Perception - Duality AI

Digital Twins for a more efficient and connected product lifecycle — from design to QA

Create accurate Digital Twins of your entire production line and process

Spot, address, and troubleshoot costly production issues before they show up on the physical line. Simulate entire process or isolated steps of significant interest.

Train 99%+ accurate computer vision models for manufacturing and QA inspection systems

Falcon’s synthetic images drive more accurate, lower cost, and faster results than real-world data.

Test downstream effects of design changes on the entire process

Immediate feedback loop from production, QA, packaging, and more.

Acquire high-fidelity, photo-real Digital Twins at scale and from existent assets

Leverage your CAD-type files, 3D asset libraries, or custom-made models.

Reuse same Digital Twins for every part of the product lifecycle

Manufacturing and QA simulation, product images, metaverse experiences, marketing materials, and beyond.

Falcon is designed to work with your process and fit into your workflow

User-friendly design and our support team help you focus on solving challenges, rather than on Falcon.

Featured Customer + Application

Manufacturing Defect Detection

User Need: Manufacturer needs a 99.9+% accurate detection rate from their ML-enabled QA system. Using real-world images to train the model has proven time and resource intensive. And short of their accuracy goal.

Falcon Workflow: Using Falcon, a high-fidelity Digital Twin of the product is generated using existing CAD files. Realistic, customizable imperfections, in vast permutations, are created on the Digital Twin surface without need for a new model for each iteration. Using Falcon's virtual camera sensors and the customizable Digital Twin, hundreds of thousands of indistinguishable-from-real images are generated in record time. After training the QA ML model with the Falcon-produced synthetic images, the manufacturer is able to surpass their QA goal on the real production line.