Methods and apparatus for the application of machine learning to radiographic images of animals

A non-transitory computer-readable storage apparatus comprising a plurality of instructions, that when executed by a processor apparatus, are configured to:

• receive a set of radiographic images captured of an animal;
• apply one or more transformations to at least a portion of the set of radiographic images captured of the animal, the application including one or more of a rotation operation, a translation operation, and a normalization operation to create a modified set of radiographic images;
• segment the modified set of radiographic images using one or more segmentation artificial intelligence engines to create a set of segmented radiographic images;
• feed the set of segmented radiographic images to respective ones of a plurality of classification artificial intelligence engines;
• output results from the plurality of classification artificial intelligence engines for the set of segmented radiographic images to an output decision engine;
• add the set of segmented radiographic images and the output results from the plurality of classification artificial intelligence engines to a training set for one or more of the plurality of classification artificial intelligence engines;
• output results from the output decision engine to a graphical user interface (GUI), the GUI including the modified set of radiographic images, a centralized radiographic image from the modified set of radiographic images, and a plurality of classifications;
• receive a first selection for one of the plurality of classifications;
• and
• highlight one or more of the modified set of radiographic images that were utilized in assessing the first selected one of the plurality of classifications.

A system for training a plurality of classification artificial intelligence engines for classification of various maladies of an animal, comprising:

• an image quality engine which receives as input, biological data as well as one or more quality control parameters;
• one or more segmentation artificial intelligence engines that are trained using one or more segmentation training sets;
• one or more classification artificial intelligence engines that are trained using one or more classification training sets;
• and
• an output decision engine that receives as input, one or more outputs from the one or more classification artificial intelligence engines;
• wherein the system is configured to:
• receive a set of radiographic images captured of the animal by the image quality engine;
• apply, by the image quality engine, one or more transformations to at least a portion of the set of radiographic images captured of the animal, the application including one or more of a rotation operation, a translation operation, and a normalization operation to create a modified set of radiographic images;
• segment the modified set of radiographic images using the one or more segmentation artificial intelligence engines to create a set of segmented radiographic images;
• feed the set of segmented radiographic images by the one or more segmentation artificial intelligence engines to respective ones of the one or more classification artificial intelligence engines;
• output results from the one or more classification artificial intelligence engines for the set of segmented radiographic images to the output decision engine;
• add the set of segmented radiographic images and the output results from the one or more classification artificial intelligence engines to the one or more classification training sets;
• and
• apply a training set assistance procedure to the output results, the application of the training set assistance procedure comprising:
• verification of the output results by a quality control group;
• subsequent to the verification, forward on the output results to a requesting doctor of veterinary medicine (DVM);
• receive clinician verification from the requesting DVM;
• verify the clinician verification from the requesting DVM;
• and
• if necessary, update the training set for the one or more classification training sets.