Abstract: Various example embodiments are described in which an anisotropic encoder-decoder convolutional neural network architecture is employed to process multiparametric magnetic resonance images for the generation of cancer predication maps. In some example embodiments, a simplified anisotropic encoder-decoder convolutional neural network architecture may include an encoder portion that is deeper than a decoder portion. In some example embodiments, simplified network architectures may be combined with test-time-augmentation in order to facilitate training and testing with a minimal number of test subjects.
Type:
Grant
Filed:
December 11, 2019
Date of Patent:
March 5, 2024
Assignee:
NOVA SCOTIA HEALTH AUTHORITY
Inventors:
Alessandro Guida, David Hoar, Peter Lee, Steve Patterson, Sharon Clarke, Chris Bowen
Abstract: Systems and methods are provided for performing automated reconstruction of a dynamic MRI dataset that is acquired without a fixed temporal resolution. On one or more image quality metrics (IQMs) are obtained by processing a subset of the acquired dataset. In one example implementation, at each stage of an iterative process, one or more IQMs of the image subset is computed, and the parameters controlling the reconstruction and/or the strategy for data combination are adjusted to provide an improved or optimal image reconstruction. Once the IQM of the image subset satisfies acceptance criteria based on an estimate of the overall temporal fidelity of the reconstruction, the full reconstruction can be performed, and the estimate of the overall temporal fidelity can be reported based on the IQM at the final iteration.
Type:
Grant
Filed:
July 30, 2021
Date of Patent:
February 28, 2023
Assignee:
NOVA SCOTIA HEALTH AUTHORITY
Inventors:
James Rioux, Nathan Murtha, Steven Beyea
Abstract: Systems and methods are provided for performing automated reconstruction of a dynamic MRI dataset that is acquired without a fixed temporal resolution. On one or more image quality metrics (IQMs) are obtained by processing a subset of the acquired dataset. In one example implementation, at each stage of an iterative process, one or more IQMs of the image subset is computed, and the parameters controlling the reconstruction and/or the strategy for data combination are adjusted to provide an improved or optimal image reconstruction. Once the IQM of the image subset satisfies acceptance criteria based on an estimate of the overall temporal fidelity of the reconstruction, the full reconstruction can be performed, and the estimate of the overall temporal fidelity can be reported based on the IQM at the final iteration.
Type:
Grant
Filed:
July 25, 2018
Date of Patent:
August 3, 2021
Assignee:
NOVA SCOTIA HEALTH AUTHORITY
Inventors:
James Rioux, Nathan Murtha, Steven Beyea
Abstract: Systems and methods are provided for performing automated reconstruction of a dynamic MRI dataset that is acquired without a fixed temporal resolution. On one or more image quality metrics (IQMs) are obtained by processing a subset of the acquired dataset. In one example implementation, at each stage of an iterative process, one or more IQMs of the image subset is computed, and the parameters controlling the reconstruction and/or the strategy for data combination are adjusted to provide an improved or optimal image reconstruction. Once the IQM of the image subset satisfies acceptance criteria based on an estimate of the overall temporal fidelity of the reconstruction, the full reconstruction can be performed, and the estimate of the overall temporal fidelity can be reported based on the IQM at the final iteration.
Type:
Application
Filed:
July 25, 2018
Publication date:
November 19, 2020
Applicant:
Nova Scotia Health Authority
Inventors:
James RIOUX, Nathan MURTHA, Steven BEYEA