Systems for detecting and identifying coincident conditions
Inventors
Ferrante, Joseph • Swart, Elliot
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12205722-B2
Publication Date
2025-01-21
Expiration Date
Abstract
A diagnosis system trains a set of machine-learned diagnosis models that are configured to receive an image of a patient and generate predictions on whether the patient has one or more health conditions. In one embodiment, the set of machine-learned models are trained to generate predictions for images that contain two or more underlying health conditions of the patient. In one instance, the symptoms for the two or more health conditions are shown as two or more overlapping skin abnormalities on the patient. By using the architectures of the set of diagnosis models described herein, the diagnosis system can generate more accurate predictions for images that contain overlapping symptoms for two or more health conditions compared to existing systems.
Core Innovation
The disclosure describes generating an indication of whether an input image shows two or more skin abnormalities. In response to a request from a client device for diagnoses, a system accesses a machine-learned model with trained weights and applies it to the image to estimate the likelihood of multiple abnormalities. The system uses that indication to determine whether multiple abnormalities are present and generates a result for the client device.
The disclosed approach addresses diagnosis of images that may contain multiple, potentially overlapping skin abnormalities, for which prediction accuracy can be improved. The result supports diagnosis pathways for images determined to show one abnormality or two or more abnormalities. The disclosure also describes a computer program product that performs the indication, determination, and result-generation functions.
Claims Coverage
The claims identify two independent claims: claim 1 covers a method, and claim 13 covers a computer program product. Their main inventive features are the trained-model indication of multiple skin abnormalities, the determination based on that indication, and the generation and delivery of a result.
Machine-learned indication of multiple skin abnormalities
Claim 1 receives a request for diagnoses of skin abnormalities in an input image, accesses a machine-learned model with trained weights, and applies the model to generate an indication representing a likelihood that the image shows two or more skin abnormalities.
Determination and result for the client device
Claim 1 generates a determination on whether the input image presents two or more skin abnormalities from the generated indication, generates a result for the request based on the determination, and provides the result to the client device.
Computer program product for generating an indication
Claim 13 covers a non-transitory computer program product with a computer-readable storage medium containing program code that receives a request, accesses a machine-learned model with trained weights, and applies the model to generate an indication representing a likelihood that the input image includes two or more skin abnormalities.
Computer program product for determining and returning a result
Claim 13 program code generates a determination on whether the input image presents two or more skin abnormalities from the indication, generates a result for the request based on the determination, and provides the result to the client device.
The two independent claims cover the same central sequence in different forms: a method and a computer program product use a trained machine-learned model to indicate the likelihood of multiple skin abnormalities, determine whether they are present, and return a result to the requesting client device.
Stated Advantages
The disclosure states that the approaches can improve prediction accuracy for overlapping symptoms.
Documented Applications
Generating an indication and diagnosis result for a client device from an input image containing skin abnormalities, including determining whether two or more abnormalities are present.
Interested in licensing this patent?