Main Article Content


Attractive reverberation imaging (MRI) is by and large progressively used to evaluate, analyse, and plan treatment for an assortment of illnesses. The capacity to picture tissue in changed differentiations as MR beat groupings in a solitary output gives important experiences to doctors, just as empowering mechanized frameworks performing downstream examination. Anyway numerous issues like restrictive output time, picture defilement, distinctive obtaining conventions, or hypersensitivities to certain difference materials may upset the way toward securing various successions for a patient. This postures difficulties to the two doctors and mechanized frameworks since reciprocal data given by the missing groupings is lost. In this paper, we propose a variation of generative ill-disposed organization (GAN) fit for utilizing repetitive data contained inside numerous accessible successions to produce at least one missing groupings for a patient sweep. The proposed network is planned as a multi-input, multi-yield network which consolidates data from all the accessible heartbeat arrangements and orchestrates the missing ones in a solitary forward pass. We exhibit and approve our technique on two mind MRI datasets each with four arrangements, and show the relevance of the proposed strategy in all the while incorporating all missing groupings in any conceivable situation where possibly one, two, or three of the four successions might be absent. We contrast our methodology and contending unimodal and multi-modular strategies, and show that we beat both quantitatively and subjectively.

Article Details

How to Cite
M. Hemakumar, J. Arivazhagan, S. Premkumar, D. Kowsalya, & S. Swetha. (2021). Missing MRI pulse sequence synthesis using multi-model generative adversarial network. International Journal of Intellectual Advancements and Research in Engineering Computations, 9(2), 25–29. Retrieved from