Get Started

Evaluation of Image Data Simulation Systems

  • Location: Fraunhofer, Stuttgart
Link to media file

Service description

Evaluation and/or improvement of Image Simulation/Generation Systems for optimization of Vision Systems: Training of models on real and generated data to determine domain gap and feasibility of Image Simulation approach for Vision AI Use Cases.

Expected results:

Test results/Feasability Report/Recommendation Report/Synthetic Training Dataset/Trained Model

Methodology:
  • identify requirements
  • use standardized synthetic data pipeline to generate data suitable for comparison with client’s pipeline
  • optionally train suitable AI models on synthetic data for comparison
  • evaluate and compare models on synthetic and real data
Target:

Vision AI Providers that want to employ better vision AI models but lack sufficient amount of data.

  • Organisation: IPA
  • Type: experiment
  • SUT/E: Vision AI Models
  • Average duration: 1-3 months
  • Contact: ira.effenberger@ipa.fraunhofer.de; andreas.frommknecht@ipa.fraunhofer.de

Enhance your manufacturing
project with AI technologies