Aligner damage prediction using machine learning

A non-transitory computer readable storage medium comprising instructions that, when executed by a processing device, cause the processing device to perform operations comprising:

• performing the following for each digital design for an orthodontic aligner of digital designs for a plurality of orthodontic aligners:
• processing data from the digital design for the orthodontic aligner by a trained machine learning model;
• outputting, by the trained machine learning model, a probability that the orthodontic aligner associated with the digital design will be damaged during manufacturing of the orthodontic aligner;
• and
• determining whether the orthodontic aligner is a high risk orthodontic aligner based on comparing the probability that the orthodontic aligner associated with the digital design will be damaged during manufacturing of the orthodontic aligner to a first probability threshold;
• determining an aligner set comprising the digital designs for the plurality of orthodontic aligners, wherein the digital designs for the plurality of orthodontic aligners are all associated with a treatment plan for a patient;
• determining whether any orthodontic aligner represented in the aligner set is a high risk orthodontic aligner;
• and
• responsive to determining that at least one orthodontic aligner represented in the aligner set is a high risk orthodontic aligner, selecting a manufacturing flow for high risk orthodontic aligners for the aligner set to reduce the probability that orthodontic aligners in the aligner set will be damaged during manufacturing.