Low dose digital tomosynthesis system and method using artificial intelligence
Inventors
Wang, Xiaohui • Sehnert, William J. • Vogelsang, Levon O. • Foos, David H. • Lin, Yuan
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11478213-B2
Publication Date
2022-10-25
Expiration Date
Abstract
A mobile radiography apparatus is configured to sparsely sample radiographic projection images to generate high resolution tomosynthesis volume images using a digital radiographic detector that is mechanically uncoupled from the x-ray source and an artificial intelligence network. The artificial intelligence network is trained to correct a volume image generated from sparsely sample projection images to generate the high resolution tomosynthesis volume images.
Core Innovation
The invention corrects image defects in an image volume reconstructed from sparsely captured tomosynthesis projection images. A convolutional neural network is trained to correct a reconstructed sparse volume image and outputs a corrected reconstructed sparse volume image of the patient anatomy.
The system captures tomosynthesis projection images within a limited preselected angular range and uses those sparsely captured projection images to reconstruct a sparse volume image. Sparse capture is described as covering between about two-thirds of a full angular range to about four-fifths of a full angular range, or as capturing about one-third to about one-fourth of a standard number of tomosynthesis projection images.
Training is described using full-vs-sparse reconstruction pairs, where full and sparse tomosynthesis volume images are reconstructed and differences between the full and sparse volumes are determined. A deep learning network is trained using these differences and then outputs a corrected reconstructed sparse volume image, with the corrected output described as a high resolution corrected tomosynthesis volume image. Geometric support is provided using radio-opaque markers to determine the source-to-detector geometry and the portable DR detector position relative to the x-ray source.
Claims Coverage
The patent includes three independent claims covering computer implemented radiographic image processing, a mobile radiography system with a mechanically uncoupled portable DR detector, and a method variant specifying a different level of sparsity by reducing the number of tomosynthesis projection images.
Neural network correction of defects in reconstructed sparse tomosynthesis volumes
Training a convolutional neural network to correct image defects in an image volume reconstructed from sparsely captured tomosynthesis projection images, providing the reconstructed sparse volume image to the trained convolutional neural network, and outputting a corrected reconstructed sparse volume image.
Sparse tomosynthesis projection capture within a limited angular range
Sparsely capturing tomosynthesis projection images of a patient anatomy within an angular range between about two-thirds of a full angular range to about four-fifths of the full angular range using at least one x-ray source and a digital radiographic (DR) detector.
Mobile radiography system with mechanically uncoupled portable DR detector for sparse tomosynthesis capture
Providing a mobile base, a support arm attached to the mobile base, an x-ray tube head comprising one or more x-ray sources attached to one end of the support arm, and a portable DR detector mechanically uncoupled from the x-ray tube head, the portable DR detector and the x-ray tube head configured to sparsely capture tomosynthesis projection images of a patient anatomy over a preselected angular range and reconstruct a sparse volume image using the sparsely captured tomosynthesis projection images, with an artificial intelligence module trained to correct the reconstructed sparse volume image and output a corrected reconstructed sparse volume image.
Different sparse projection reduction by capturing about one-third to about one-fourth of a standard number
Sparsely capturing tomosynthesis projection images of a patient anatomy at about one-third to about one-fourth of a standard number of tomosynthesis projection images using at least one x-ray source and a digital radiographic (DR) detector, reconstructing a sparse volume image, providing the reconstructed sparse volume image to a trained convolutional neural network, and outputting a corrected reconstructed sparse volume image.
Across the independent claims, the inventive coverage centers on training a convolutional neural network or an artificial intelligence module to correct defects in reconstructed sparse tomosynthesis volume images, with sparsity achieved through limited angular range or reduced number of tomosynthesis projection images, and a mobile radiography arrangement with a mechanically uncoupled portable DR detector configured for sparse capture.
Stated Advantages
Corrected reconstructed sparse volume images are output by the trained neural network.
Sparse capture is performed using a preselected angular range or using about one-third to about one-fourth of a standard number of tomosynthesis projection images.
Geometric determination is provided using radio-opaque markers to determine source-to-detector geometry, including portable DR detector position relative to the x-ray source.
Full-vs-sparse reconstruction differences are used to train the convolutional neural network for defect correction.
Documented Applications
No documented applications found
Interested in licensing this patent?