Abstract:
MRI is increasingly used to assess, diagnose, and plan treatment for various diseases. Physicians and automated systems can benefit from MR pulse sequences that show tissue in different contrasts in a single scan.
However, scan time, image corruption, acquisition protocols, and patient allergies to contrast materials may make it difficult to acquire multiple sequences. The missing sequences’ complementary information challenges physicians and automated systems.
We propose a generative adversarial network (GAN) that uses redundant information from multiple sequences to generate one or more missing sequences for a patient scan. The proposed multi-input, multi-output network combines information from all pulse sequences, implicitly infers missing sequences, and synthesizes them in a single forward pass.
We demonstrate and validate our method on two brain MRI datasets each with four sequences and show that it can synthesize all missing sequences in any scenario where one, two, or three sequences are missing. We outperform unimodal and multimodal methods quantitatively and qualitatively.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
