Utilize Machine Learning to Diagnose Specific Conditions

Splenic Mass

The images are processed in order to reveal additional features from the original image.  The features that correlate with malignancies, normal and benign tissues are used to build a machine learning model.  The machine learning model can then be use to generate a score to identify future images as normal, benign or malignant.  We currently building models to identify the following conditions:

Splenic Mass

  • Source Image:  Ultrasound
  • Status:  99% separation between benign and malignant lesions
  • Next Steps:  Test more types of benign (4 types tested) and malignant (3 types tested) lesions

 

Stifel Joint

  • Source Image:  Radiograph
  • Status:  Beta Version Operational, 100% separation between normal and torn ACL
  • Next Steps:  Test more images and conditions