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

A method for identification of various physiological conditions of an animal using machine learning, the method comprising:

• receiving one or more radiographic images captured of the animal;
• receiving non-imaging biological data for the animal;
• segmenting the received one or more radiographic images captured of the animal using one or more segmentation artificial intelligence engines to create a set of segmented radiographic images, each of the set of segmented radiographic images corresponding to a specific anatomical area within the animal, the segmenting of the received one or more radiographic images captured of the animal comprises segregating a given organ and/or a given skeletal structure from remaining portions of the received one or more radiographic images captured of the animal;
• providing the set of segmented radiographic images to respective ones of a plurality of classification artificial intelligence engines;
• providing the received non-imaging biological data for the animal to one or more of the plurality of classification artificial intelligence engines;
• outputting results from the plurality of classification artificial intelligence engines for the set of segmented radiographic images and the non-imaging biological data to an output decision engine, the outputting of the results from the plurality of classification artificial intelligence engines comprises outputting either a normal condition or an abnormal condition along with a confidence level associated with the normal condition or the abnormal condition;
• providing recommended courses of action, using the output decision engine, based on the output results from the plurality of classification artificial intelligence engines;
• analyzing metadata associated with the received one or more radiographic images captured of the animal;
• comparing the analyzed metadata with the output results in order to determine misapplied parameters associated with a subset of the plurality of classification artificial intelligence engines;
• and
• selectively discarding the output results associated with the misapplied parameters.