Autonomous healthcare visual system for real-time prevention, diagnostics, treatments and rehabilitation
AssigneesPlanned Systems International, Inc., National Society of Medical Scientists
InventorsJune Lee, Thomas J. Berti, Yan Li
Life of this patent
Estimating the term…
Abstract
Systems, methods, and other embodiments relate to autonomous clinical screening and assessment of a patient using a reinforcement learning framework. In at least one approach, a method includes determining, using a control model, a focus point for acquiring imaging data about the patient. The method includes controlling an imaging device according to the focus point to acquire the imaging data. The method includes analyzing the imaging data using a recognition model that outputs a result about a condition of the patient. The method includes providing the result.
Claims20
A screening system for autonomous clinical screening and assessment of a patient, comprising:
- one or more processors;
- and
- a memory communicably coupled to the one or more processors and storing instructions that, when executed by the one or more processors, cause the one or more processors to:
- determine, using a control model comprising a deep reinforcement learning (DRL) model, a focus point for acquiring imaging data about the patient by analyzing a condition of the patient from a prior acquisition of the imaging data to identify a location on the patient for the focus point and attributes of an imaging device;
- control the imaging device according to the focus point associated with the condition to acquire the imaging data by controlling one or more actuators and motors to move the imaging device into a position to image the patient according to the focus point and to focus and zoom the imaging device to acquire the imaging data;
- analyze the imaging data using a recognition model that outputs a result about the condition of the patient, the recognition model comprising a machine learning model selected from a group consisting of a convolutional neural network (CNN), a recurrent neural network (RNN), a support vector machine (SVM), and a random forest;
- and
- provide the result.
A non-transitory computer-readable medium for autonomous clinical screening and assessment of a patient and storing instructions that, when executed by one or more processors, cause the one or more processors to:
- determine, using a control model comprising a deep reinforcement learning (DRL) model, a focus point for acquiring imaging data about the patient by analyzing a condition of the patient from a prior acquisition of the imaging data to identify a location on the patient for the focus point and attributes of an imaging device;
- control the imaging device according to the focus point associated with the condition to acquire the imaging data by controlling one or more actuators and motors to move the imaging device into a position to image the patient according to the focus point and to focus and zoom the imaging device to acquire the imaging data;
- analyze the imaging data using a recognition model that outputs a result about the condition of the patient, the recognition model comprising a machine learning model selected from a group consisting of a convolutional neural network (CNN), a recurrent neural network (RNN), a support vector machine (SVM), and a random forest;
- and
- provide the result.
A method for autonomous clinical screening and assessment of a patient, comprising:
- determining, using a control model comprising a deep reinforcement learning (DRL) model, a focus point for acquiring imaging data about the patient by analyzing a condition of the patient from a prior acquisition of the imaging data to identify a location on the patient for the focus point and attributes of an imaging device;
- controlling the imaging device according to the focus point associated with the condition to acquire the imaging data by controlling one or more actuators and motors to move the imaging device into a position to image the patient according to the focus point and to focus and zoom the imaging device to acquire the imaging data;
- analyzing the imaging data using a recognition model that outputs a result about the condition of the patient, the recognition model comprising a machine learning model selected from a group consisting of a convolutional neural network (CNN), a recurrent neural network (RNN), a support vector machine (SVM), and a random forest;
- and
- providing the result.
Dependencies are inferred from the claim text.