Systems and methods for automated physiological parameter estimation from ultrasound image sequences
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12089988-B2
Publication Date
2024-09-17
Expiration Date
Abstract
Systems and methods for automated physiological parameter estimation from ultrasound image sequences are provided. An ultrasound system includes an ultrasound imaging device configured to acquire a sequence of ultrasound images of a patient. An anatomical structure recognition module includes processing circuitry configured to receive the acquired sequence of ultrasound images from the ultrasound imaging device, and automatically recognize an anatomical structure in the received sequence of ultrasound images. A physiological parameters estimation module includes processing circuitry configured to automatically estimate one or more physiological parameters associated with the recognized anatomical structure.
Core Innovation
The invention relates to an ultrasound system and corresponding method that acquires a sequence of ultrasound images of a patient and automatically recognizes an anatomical structure in the sequence. The system includes an anatomical structure recognition module with processing circuitry configured to receive the sequence and perform the automatic recognition. The approach determines a quality of the sequence of ultrasound images using an ultrasound image quality module with processing circuitry.
The system further includes a physiological parameters estimation module with processing circuitry configured to automatically estimate one or more physiological parameters associated with the recognized anatomical structure. The anatomical structure recognition module and the physiological parameters estimation module are implemented using one or more machine learning models. During inference, the machine learning models receive the entire input ultrasound image sequence as a whole and automatically operate on the entire input sequence to output an inference estimating the one or more physiological parameters.
The inference is based at least in part on prior knowledge imposed on the machine learning models during training. The prior knowledge includes spatiotemporal prior knowledge associated with the anatomical structure and indicative of a periodic relationship between the anatomical structure and the sequence of ultrasound images, supporting temporal consistency across the sequence. The physiological parameter estimation is based at least in part on the quality of the sequence of ultrasound images.
Claims Coverage
The provided independent claims are clm-00001 and clm-00012. Both claims share four inventive features: ultrasound sequence acquisition, anatomical structure recognition, ultrasound image quality determination, and physiological parameter estimation using machine learning models that perform inference on the entire sequence as a whole with spatiotemporal periodic prior knowledge.
Sequence-based anatomical structure recognition and physiological parameter estimation
An anatomical structure recognition module/circuitry automatically recognizes an anatomical structure in a sequence of ultrasound images, and a physiological parameters estimation module/circuitry automatically estimates one or more physiological parameters associated with the anatomical structure.
Machine learning inference over the entire ultrasound image sequence
One or more machine learning models receive the sequence as an input ultrasound image sequence and, during inference, automatically operate on the entire input ultrasound image sequence as a whole to output an inference that estimates the one or more physiological parameters associated with the anatomical structure.
Spatiotemporal periodic prior knowledge
The inference process is based at least in part on prior knowledge imposed on the one or more machine learning models during training, wherein the prior knowledge includes spatiotemporal prior knowledge associated with the anatomical structure and indicative of a periodic relationship between the anatomical structure and the sequence of ultrasound images.
Physiological parameter estimation based on sequence quality
The one or more machine learning models are configured to automatically estimate the one or more physiological parameters based at least in part on the quality of the sequence of ultrasound images, where the system includes an ultrasound image quality module/circuitry configured to determine the quality of the sequence.
Across clm-00001 and clm-00012, the claims converge on a machine-learning workflow for ultrasound: recognize an anatomical structure in an image sequence, determine sequence quality, and estimate one or more physiological parameters using models that run inference on the entire sequence as a whole, with spatiotemporal periodic prior knowledge and quality-based estimation.
Stated Advantages
Documented Applications
No documented applications found
Interested in licensing this patent?