Methods and computing system for processing ultrasound image to determine health of subdermal tissue

#The present invention has application in image processing, machine learning, and sonography.

#The present disclosure relates to a computing system, a method, and instructions on a non-transitory computer-readable medium for processing ultrasound images. The computing system may receive an ultrasound image that includes image data representing subdermal tissue, and apply a trained neural network to the ultrasound image to generate at least one of: a pliability parameter value which indicates pliability of the subdermal tissue, one or more regions of the ultrasound image that represents damaged tissue, or an indication of presence and/or location of an entrapped nerve.

#The foregoing and other features and advantages of the invention will be apparent from the following description of embodiments hereof as illustrated in the accompanying drawings. The accompanying drawings, which are incorporated herein and form a part of the specification, further serve to explain the principles of the invention and to enable a person skilled in the pertinent art to make and use the invention. The drawings are not to scale.

#FIG. 1 provides a block diagram of a system for processing ultrasound imaging data to assess health of connective tissue.

#FIG. 2 provides a block diagram of an image processing module configured to process an ultrasound image with a trained neural network.

#The following detailed description is merely exemplary in nature and is not intended to limit the invention or the application and uses of the invention. Furthermore, there is no intention to be bound by any expressed or implied theory presented in the preceding technical field, background, brief summary or the following detailed description. Specific embodiments of the present invention are now described with reference to the figures, wherein like reference numbers indicate identical or functionally similar elements.

#One aspect of the embodiments herein relates to a computing system for processing an ultrasound image using a trained neural network (or other form of machine learning) to assess health of connective tissue (or, more generally, mesoderm-derived tissue) captured or otherwise represented by the ultrasound image. The ultrasound image may be derived from performing ultrasound scanning to sense subdermal tissue that is, e.g., within 2-5 cm of the skin surface. In an embodiment, the ultrasound image may represent at least a connective tissue layer within that depth, wherein the connective tissue layer may be located between the skin and muscle. In some instances, the ultrasound image may represent the connective tissue layer as well as underlying layers (e.g., deeper layers), such as a muscle layer and/or a bone layer.

#In an embodiment, the computing system may receive an ultrasound image that represents subdermal tissue, and may use a trained neural network to assess overall health of tissue represented by the ultrasound image. The subdermal tissue may include, e.g., a layer of connective tissue. In some cases, the neural network may be used to generate a value or range of values which indicate the health of the tissue represented by the ultrasound image. For instance, the computing system may use the trained neural network to generate a value of a pliability parameter that indicates pliability or mobility of tissue represented by the ultrasound image, wherein the value of the pliability parameter may be an inverse of a value of a stiffness parameter that indicates stiffness of the connective tissue layer. In other words, the pliability parameter may indicate how much mobility or pliability there is in the tissue represented by the ultrasound image, or indicate how little stiffness there is in the tissue. The value of the pliability parameter may be referred to as a pliability parameter value or a pliability score, and may be the inverse of a stiffness parameter value or stiffness score. In a more specific example, the pliability score may indicate tissue fluidity. In such an example, a higher pliability score may indicate that more soft tissue and less hard tissue are found or represented in the ultrasound image, while a lower pliability score may indicate that more hard tissue and less soft tissue are found in or represented in the ultrasound image. Thus, one aspect of the embodiments herein relates to a computing system that is configured to use a trained neural network and an ultrasound image(s) to generate a metric which indicates health, or more specifically pliability, of connective tissue represented in the ultrasound image(s). This metric may be a pliability parameter value (also referred to as a pliability score), which may be used to measure pliability of connective tissue.