Probabilistic medical image imputation via deep adversarial learning

Raad, Ragheb; Patel, Dhruv; Hsu, Chiao-Chih; Kothapalli, Vijay; Ray, Deep; Varghese, Bino; Hwang, Darryl; Gill, Inderbir S.; Duddalwar, Vinay; Oberai, Assad A.

doi:10.1007/s00366-022-01712-8

Cited by 5 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We remark that in an earlier work [25] we presented a probabilistic method for imputing CECT images where we utilized GANs to learn the prior distribution [23,24] of a complete sequence of CECT images. This prior was combined with a likelihood term driven by a measured incomplete sequence to set up a Bayesian estimate for the probability distribution of the complete sequence.…”

Section: Precontrast Nephrographic Excretorymentioning

confidence: 99%

Conditional Generative Learning for Medical Image Imputation

Raad

Ray

Varghese

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Image imputation refers to the task of generating a type of medical image given images of another type. This task becomes challenging when the difference between the available images, and the image to be imputed is large. In this manuscript, one such application, derived from the dynamic contrast enhanced computed tomography (CECT) imaging of the kidneys, is considered: given an incomplete sequence of three CECT images, we are required to the impute the missing image. This task is posed as one of probabilistic inference and a generative algorithm to generate samples of the imputed image, conditioned on the available images, is developed, trained, and tested. The output of this algorithm is the best guess of the imputed image, and a pixel-wise image of variance in the imputation. It is demonstrated that this best guess is more accurate than those generated by other, deterministic deep-learning based algorithms, including ones which utilize additional information and more complex loss terms. It is also shown the pixel-wise variance image, which quantifies the confidence in the reconstruction, can be used to determine whether the result of the imputation meets a specified accuracy threshold and is therefore appropriate for a downstream task.

show abstract

Section: Precontrast Nephrographic Excretorymentioning

confidence: 99%

Conditional Generative Learning for Medical Image Imputation

Raad

Ray

Varghese

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We remark that in an earlier work 26 we presented a probabilistic method for imputing CECT images where we utilized GANs to learn the prior distribution 27 , 28 of a complete sequence of CECT images. This prior was combined with a likelihood term driven by a measured incomplete sequence to set up a Bayesian estimate for the probability distribution of the complete sequence.…”

Section: Introductionmentioning

confidence: 99%

Conditional generative learning for medical image imputation

Raad,

Ray,

Varghese

et al. 2024

Sci Rep

Self Cite

View full text Add to dashboard Cite

Image imputation refers to the task of generating a type of medical image given images of another type. This task becomes challenging when the difference between the available images, and the image to be imputed is large. In this manuscript, one such application is considered. It is derived from the dynamic contrast enhanced computed tomography (CECT) imaging of the kidneys: given an incomplete sequence of three CECT images, we are required to impute the missing image. This task is posed as one of probabilistic inference and a generative algorithm to generate samples of the imputed image, conditioned on the available images, is developed, trained, and tested. The output of this algorithm is the “best guess” of the imputed image, and a pixel-wise image of variance in the imputation. It is demonstrated that this best guess is more accurate than those generated by other, deterministic deep-learning based algorithms, including ones which utilize additional information and more complex loss terms. It is also shown that the pixel-wise variance image, which quantifies the confidence in the reconstruction, can be used to determine whether the result of the imputation meets a specified accuracy threshold and is therefore appropriate for a downstream task.

show abstract

“…For instance, ML methods can enhance prediction of scoliosis progression given longitudinal data for patients at risk by integrating these data with the modeling of bone biomechanics and growth [15]. ML is suitable for posing inverse problems such as the identification of cardiac activation maps [10], or imputation of missing longitudinal data in an image sequence [9]. Undoubtedly, work in ML that specifically fuses image analysis and biophysics modeling will continue to play a central role in computational medicine.…”

mentioning

confidence: 99%