About this Abstract |
Meeting |
2021 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2021)
|
Symposium
|
Special Session
|
Presentation Title |
Automated Detection of Part Quality during Two Photon Lithography via Deep Learning |
Author(s) |
Brian Giera, Xian Lee, Sourabh K Saha, Soumik Sarkar |
On-Site Speaker (Planned) |
Xian Lee |
Abstract Scope |
Two photon lithography (TPL) requires exhaustive parameter optimization process to identify a suitable light dosage that sufficiently cures a photo-reactive polymer. This involves experts analyzing routinely-collected video of every part. This laborious process limits the adoption of TPL as an industrial-scale additive manufacturing technology. Here, we solve this critical issue via deep learning, leveraging TPL video to train algorithms that automatically predict part quality at 95% accuracy in milliseconds. We demonstrate a general procedure to curate labeled TPL video data set that we publicly released. We implement and evaluate several deep learning model architectures and find we can reliably classify both the state of photo-polymerization and overall part quality, eliminating the need for human oversight during the parameter optimization step. Furthermore, these algorithms can be used for real-time part quality monitoring.
This work was performed under the auspices of the U.S. Department of Energy by LLNL under Contract DE-AC52-07NA27344. |
Proceedings Inclusion? |
Definite: Post-meeting proceedings |