Defect detection of molds for orthodontic aligners

Abstract

Implementations describe systems and methods for machine based defect detection of molds for orthodontic aligners. A method of one embodiment of the disclosure includes receiving, by a processor, a digital representation of the fabricated mold for the orthodontic aligner, the digital representation having been generated based on imaging of the fabricated mold. The method further includes analyzing, by the processor, the digital representation of the fabricated mold to identify a quality-related property of the fabricated mold. The method further includes determining, based on the quality-related property, that the fabricated mold comprises a defect. The method further includes classifying, by the processor, the fabricated mold as defective based on determining that the fabricated mold comprises the defect.

What is claimed is: 
     
         1 . A method for analyzing a quality of a fabricated mold for an orthodontic aligner, the method comprising:
 receiving, by a processor, a digital representation of the fabricated mold for the orthodontic aligner, the digital representation having been generated based on imaging of the fabricated mold;   analyzing, by the processor, the digital representation of the fabricated mold to identify a quality-related property of the fabricated mold;   determining, based on the quality-related property, that the fabricated mold comprises a defect; and   classifying, by the processor, the fabricated mold as defective based on determining that the fabricated mold comprises the defect.   
     
     
         2 . The method of  claim 1 , further comprising:
 determining, based on the defect, whether the fabricated mold is suitable for use in forming the orthodontic aligner.   
     
     
         3 . The method of  claim 1 , wherein the digital representation of the fabricated mold comprises one or more two-dimensional images corresponding to one or more views of the fabricated mold. 
     
     
         4 . The method of  claim 3 , further comprising:
 comparing, by the processor, an image from among the one or more two-dimensional images of the fabricated mold with a projection of a virtual three-dimensional (3D) model associated with the fabricated mold.   
     
     
         5 . The method of  claim 3 , wherein the one or more two-dimensional images comprise at least a top down image, a side image and a front image. 
     
     
         6 . The method of  claim 1 , wherein one or more features of the fabricated mold are enhanced via illumination to facilitate capture of the one or more features when generating the digital representation of the fabricated mold. 
     
     
         7 . The method of  claim 1 , further comprising:
 providing an illumination of the fabricated mold using a light source arrangement; and   generating the digital representation of the fabricated mold using one or more imaging devices, wherein the digital representation of the fabricated mold comprises a plurality of images of the fabricated mold, each image of the plurality of images depicting a distinct region of the fabricated mold.   
     
     
         8 . The method of  claim 1 , further comprising:
 processing the digital representation of the fabricated mold using a machine learning model trained to identify one or more types of defects of three-dimensionally printed molds, wherein an output of the machine learning model comprises, for each type of defect, a probability that the digital representation comprises a defect of that type of defect.   
     
     
         9 . The method of  claim 1 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises an internal volume defect within an internal volume of the fabricated mold. 
     
     
         10 . The method of  claim 1 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises a surface defect on a surface of the fabricated mold. 
     
     
         11 . The method of  claim 1 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises an interface defect at an interface of an internal volume of the fabricated mold and a surface of the fabricated mold. 
     
     
         12 . The method of  claim 1 , wherein the defect comprises at least one of a break or a hole in a wall of the fabricated mold. 
     
     
         13 . A system for analyzing a quality of a fabricated mold for an orthodontic aligner, the system comprising:
 a processor and memory coupled to the processor and storing instructions that, when executed by the processor, cause the processor to:   receive a digital representation of the fabricated mold for the orthodontic aligner, the digital representation having been generated based on imaging of the fabricated mold;   analyze the digital representation of the fabricated mold to identify a quality-related property of the fabricated mold;   determine, based on the quality-related property, that the fabricated mold comprises a defect; and   classify the fabricated mold as having the defect.   
     
     
         14 . The system of  claim 13 , wherein the processor is further to:
 determine, based on the defect, whether the fabricated mold is suitable for use in forming the orthodontic aligner.   
     
     
         15 . The system of  claim 13 , wherein the digital representation of the fabricated mold comprises one or more two-dimensional images corresponding to one or more views of the fabricated mold. 
     
     
         16 . The system of  claim 15 , wherein the processor is further to:
 compare an image from among the one or more two-dimensional images of the fabricated mold with a projection of a virtual three-dimensional (3D) model associated with the fabricated mold.   
     
     
         17 . The system of  claim 13 , wherein one or more features of the fabricated mold are enhanced via illumination to facilitate capture of the one or more features when generating the digital representation of the fabricated mold. 
     
     
         18 . The system of  claim 13 , wherein the processor is further to:
 process the digital representation of the fabricated mold using a machine learning model trained to identify one or more types of defects of three-dimensionally printed molds, wherein an output of the machine learning model comprises, for each type of defect, a probability that the digital representation comprises a defect of that type of defect.   
     
     
         19 . The system of  claim 13 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises an internal volume defect within an internal volume of the fabricated mold. 
     
     
         20 . The system of  claim 13 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises a surface defect on a surface of the fabricated mold. 
     
     
         21 . The system of  claim 13 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises an interface defect at an interface of an internal volume of the fabricated mold and a surface of the fabricated mold. 
     
     
         22 . The system of  claim 13 , wherein the defect comprises at least one of a break or a hole in a wall of the fabricated mold. 
     
     
         23 . A quality control system comprising:
 an image capture device configured to generate a digital representation of a fabricated mold for an orthodontic aligner;   a light source configured to illuminate the fabricated mold in a manner that enhances an image quality of the digital representation of the fabricated mold; and   a computing device to:
 analyze a quality-related property of the digital representation of the fabricated mold; 
 determine, based on the quality-related property, that the fabricated mold comprises a defect; and 
 classify the fabricated mold as defective based on determining that the fabricated mold comprises the defect. 
   
     
     
         24 . A non-transitory computer readable medium comprising instructions that, when executed by a processor, cause the processor to perform operations comprising:
 receiving, by a processor, a digital representation of a fabricated mold for an orthodontic aligner, the digital representation having been generated based on imaging of the fabricated mold;   analyzing, by the processor, the digital representation of the fabricated mold to identify a quality-related property of the fabricated mold;   determining, based on the quality-related property, that the fabricated mold comprises a defect; and   classifying, by the processor, the fabricated mold as defective based on determining that the fabricated mold comprises the defect.   
     
     
         25 . The non-transitory computer readable medium of  claim 24 , the operations further comprising:
 determining, based on the defect, whether the fabricated mold is suitable for use in forming the orthodontic aligner.   
     
     
         26 . The non-transitory computer readable medium of  claim 24 , wherein the digital representation of the fabricated mold comprises one or more two-dimensional images corresponding to one or more views of the fabricated mold. 
     
     
         27 . The non-transitory computer readable medium of  claim 26 , the operations further comprising:
 comparing, by the processor, an image from among the one or more two-dimensional images of the fabricated mold with a projection of a virtual three-dimensional (3D) model associated with the fabricated mold.   
     
     
         28 . The non-transitory computer readable medium of  claim 24 , wherein one or more features of the fabricated mold are enhanced via illumination to facilitate capture of the one or more features when generating the digital representation of the fabricated mold. 
     
     
         29 . The non-transitory computer readable medium of  claim 24 , the operations further comprising:
 processing the digital representation of the fabricated mold using a machine learning model trained to identify one or more types of defects of three-dimensionally printed molds, wherein an output of the machine learning model comprises, for each type of defect, a probability that the digital representation comprises a defect of that type of defect.   
     
     
         30 . The non-transitory computer readable medium of  claim 24 , wherein the fabricated mold is a three-dimensional printed mold, and wherein the defect comprises at least one of an internal volume defect within an internal volume of the fabricated mold, a surface defect on a surface of the fabricated mold, or an interface defect at an interface of the internal volume of the fabricated mold and the surface of the fabricated mold.