Real-time Analytics
Delta Electronics needed to reconfigure complex production lines and train inspection models quickly, but manual position-switch changes, fragmented 3D tools, and slow hand-labeled image collection were driving downtime and limiting flexibility.
Reported impact
Delta cut the time to generate 1,000 labeled training images from two days to 10 minutes, increased synthetic-data generation by more than 100x, achieved about 90 percent object-detection accuracy with synthetic data, and reduced production downtime during line reconfiguration.
Delta built a manufacturing digital-twin workflow on NVIDIA Omniverse, OpenUSD, Isaac Sim, and Omniverse Replicator so line design, simulation, and computer-vision training could happen before production changes hit the floor. The company aggregates 3D assets from tools such as Autodesk 3ds Max, FlexSim, and Visual Components into a unified USD stage, where teams collaborate in one environment instead of shuttling files between incompatible applications. Once the digital twin is assembled, Delta runs what-if simulations to validate layout changes and uses Replicator to create photorealistic synthetic images for tray and product detection models. Those images then feed vision models and Isaac Sim camera simulations for automated optical inspection and production-line optimization. The result is a single pipeline for design, validation, and AI training that reduces reconfiguration downtime and makes high-accuracy inspection models faster to build.
3D production-line assets are brought into OpenUSD and synchronized in Omniverse, where Delta runs simulations to validate layouts and failure points. Replicator then generates synthetic inspection images that feed computer-vision models and Isaac Sim camera tests before the final configurations are pushed to the physical line.
Delta built a physically accurate digital twin of their production lines.
Now, in just 10 minutes, they can generate the same amount of labeled images that would take two days manually.
The team also achieves a 90% accuracy with synthetic data, similar to real data.
Using Isaac Sim and Omniverse Replicator has allowed us to significantly increase our synthetic data generation capabilities by over 100X.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Aug 25, 2026 Siemens Siemens needed to help industrial teams overcome fragmented engineering tools, labor shortages, and slow design and maintenance workflows so they could design, simulate, and operate complex products and factories faster… 10%–15% reductions in capital expenditure NVIDIA CUDA · NVIDIA DGX B200 · NVIDIA DGX H200 NVIDIA Sample Aug 12, 2026 Pegatron Pegatron needed to improve factory planning, quality control, and operational efficiency across large manufacturing sites, but manual line design, fragmented agent development, scarce defect data, and labor-intensive… 10% decrease in operational effort NVIDIA AI Enterprise · NVIDIA Cosmos · NVIDIA DGX NVIDIA Sample Jul 21, 2026 Wistron Corporation Wistron needed to scale production of NVIDIA AI infrastructure while improving quality, energy efficiency, and worker training across global plants, but traditional CFD studies, sparse defect data, and reactive factory… 10% energy reduction Apple Vision Pro · Cadence Reality Digital Twin Platform · NVIDIA Cosmos 