The company Backflip AI has presented a new AI-supported model that automatically generates digital models from 3D scans and optimizes them for 3D printing or CNC manufacturing. This allows spare parts to be manufactured faster and more efficiently.
In addition to this AI technology, Backflip AI has developed two new tools that can be integrated into existing workflows. A plug-in for SOLIDWORKS enables the direct conversion of scans into native CAD models, while a web application converts 3D scans into parametric files. These solutions greatly simplify a previously complex process and reduce the time required to create replacement parts to just a few minutes.
Third-party suppliers install many production lines without CAD documentation, so digital blueprints for important components are missing. Some components are no longer available or have never been digitally recorded. In the aviation industry in particular, machines are in use that date back to times when digital models did not yet exist. Backflip AI’s technology solves this problem by converting existing components into 3D printable models in just a few minutes.
Another obstacle to date has been the complexity of 3D scans. Many scanners capture fine surface details that are difficult to convert into production-ready models using traditional CAD programs. Backflip AI uses neural networks to generate clean geometries from point clouds that can be processed directly in CAD software.
The underlying AI has been trained with over 100 million geometries and, according to Backflip AI, sets new standards in efficiency and accuracy. The technology offers 60 times the training efficiency, 10 times the inference speed and 100 times the spatial resolution compared to other methods.
The two new tools will soon be available to the general public. Interested users can join a waiting list on the company’s website.
Subscribe to our Newsletter
3DPresso is a weekly newsletter that links to the most exciting global stories from the 3D printing and additive manufacturing industry.