Project SynthNet
Year 2021
Categories CAD / graphics / shape retrievalImage GenerationDeep Learning

SynthNet - Visual Search on Synthetic Images from real-world CAD data

The goal of this project is to explore generative neural approaches to create synthetic image data needed to build and optimize a visual search index. The SynthNet solution approach consists of synthetically generating all required comparison images from a real-world industry grade 3D CAD data and using them for object identification. For each component, a set of views of an object is automatically generated synthetically under different lighting and material conditions.


CAD Models to Real-World Images: A Practical Approach to Unsupervised Domain Adaptation in Industrial Object Classificatio. Dennis Ritter, Mike Hemberger, Marc Hönig, Volker Stopp, Eric Rodner, Kristian Hildebrand. Adapting to Change: Reliable Learning Across Domains - ECML-PKDD 2023 International Workshop.source and data


. .
Computer Graphics and Intelligent Interactive Systems