Method and system for material decomposition in dual-or multiple-energy x-ray based imaging

Inventors

Peng, Yu

Assignees

Curvebeam AI Ltd

Abstract

A method and system for generating material decomposition images from plural-energy x-ray based imaging, the method comprising: modelling spatial relationships and spectral relationships among the plurality of images by learning features from the plurality of images in combination and one or more of the plurality of images individually with a deep learning neural network; generating one or more basis material images employing the spatial relationships and the spectral relationships; and generating one or more material specific or material decomposition images from the basis material images. The neural network has an encoder-decoder structure and includes a plurality of encoder branches; each of one or more of the plurality of encoder branches encodes two or more images of the plurality of images in combination; and each of one or more of the plurality of encoder branches encodes a respective individual image of the plurality of images.

Core Innovation

A deep learning encoder-decoder neural network models spatial relationships and spectral relationships among a plurality of images obtained with plural-energy x-ray based imaging. The plurality of images correspond to respective energies of the plural-energy x-ray based imaging, and the modelling learns features from the plurality of images in combination and one or more of the plurality of images individually.

The spatial relationships and the spectral relationships are used to generate one or more basis material images. From the one or more basis material images, the system generates one or more material specific or material decomposition images. At least one of the generated material specific or material decomposition images is used to diagnose, identify or monitor a pathology or disease in a subject or patient obtained with plural-energy x-ray based imaging.

The described approach optionally generates one or more metal artefact images or beam hardening reduction images. Image quality of at least one of the generated material specific or material decomposition images is improved using the one or more metal artefact or beam hardening reduction images. The encoder-decoder architecture employs multiple encoder branches, including branches that encode two or more images in combination and branches that encode respective individual images.

Claims Coverage

The document provides multiple independent claims covering four aspects: a diagnostic/monitoring method, a diagnostic/monitoring system, an image-quality improvement method, and an image-quality improvement system, plus computer-readable medium implementations. Across these independents, the claims share inventive features consisting of modelling spatial and spectral relationships with a deep learning encoder-decoder network to generate basis material images, then generating material specific/material decomposition images for diagnosis or generating metal artefact/beam hardening reduction images for image quality improvement.

Across the independent claims, the core claim coverage is the same modelling-and-generation pipeline: spatial and spectral relationships among plural-energy x-ray images are learned with a deep learning encoder-decoder network, basis material images are generated from the learned relationships, and material specific/material decomposition images are generated from the basis images for diagnosis, identification or monitoring. In the image-quality improvement independents, metal artefact and/or beam hardening reduction images are additionally generated and used to improve image quality of at least one material-specific/material-decomposition image.

Stated Advantages

Documented Applications