TY - JOUR
T1 - Assessing the Capacity of a Denoising Diffusion Probabilistic Model to Reproduce Spatial Context
AU - Deshpande, Rucha
AU - Özbey, Muzaffer
AU - Li, Hua
AU - Anastasio, Mark A.
AU - Brooks, Frank J.
N1 - Publisher Copyright:
© 1982-2012 IEEE.
PY - 2024
Y1 - 2024
N2 - Diffusion models have emerged as a popular family of deep generative models (DGMs). In the literature, it has been claimed that one class of diffusion models—denoising diffusion probabilistic models (DDPMs)—demonstrate superior image synthesis performance as compared to generative adversarial networks (GANs). To date, these claims have been evaluated using either ensemble-based methods designed for natural images, or conventional measures of image quality such as structural similarity. However, there remains an important need to understand the extent to which DDPMs can reliably learn medical imaging domain-relevant information, which is referred to as ‘spatial context’ in this work. To address this, a systematic assessment of the ability of DDPMs to learn spatial context relevant to medical imaging applications is reported for the first time. A key aspect of the studies is the use of stochastic context models (SCMs) to produce training data. In this way, the ability of the DDPMs to reliably reproduce spatial context can be quantitatively assessed by use of post-hoc image analyses. Error-rates in DDPM-generated ensembles are reported, and compared to those corresponding to other modern DGMs. The studies reveal new and important insights regarding the capacity of DDPMs to learn spatial context. Notably, the results demonstrate that DDPMs hold significant capacity for generating contextually correct images that are ‘interpolated’ between training samples, which may benefit data-augmentation tasks in ways that GANs cannot.
AB - Diffusion models have emerged as a popular family of deep generative models (DGMs). In the literature, it has been claimed that one class of diffusion models—denoising diffusion probabilistic models (DDPMs)—demonstrate superior image synthesis performance as compared to generative adversarial networks (GANs). To date, these claims have been evaluated using either ensemble-based methods designed for natural images, or conventional measures of image quality such as structural similarity. However, there remains an important need to understand the extent to which DDPMs can reliably learn medical imaging domain-relevant information, which is referred to as ‘spatial context’ in this work. To address this, a systematic assessment of the ability of DDPMs to learn spatial context relevant to medical imaging applications is reported for the first time. A key aspect of the studies is the use of stochastic context models (SCMs) to produce training data. In this way, the ability of the DDPMs to reliably reproduce spatial context can be quantitatively assessed by use of post-hoc image analyses. Error-rates in DDPM-generated ensembles are reported, and compared to those corresponding to other modern DGMs. The studies reveal new and important insights regarding the capacity of DDPMs to learn spatial context. Notably, the results demonstrate that DDPMs hold significant capacity for generating contextually correct images that are ‘interpolated’ between training samples, which may benefit data-augmentation tasks in ways that GANs cannot.
KW - Denoising diffusion probabilistic models
KW - deep generative model evaluation
KW - medical image synthesis
KW - stochastic context models
KW - stochastic object model
UR - http://www.scopus.com/inward/record.url?scp=85196068379&partnerID=8YFLogxK
U2 - 10.1109/TMI.2024.3414931
DO - 10.1109/TMI.2024.3414931
M3 - Article
C2 - 38875086
AN - SCOPUS:85196068379
SN - 0278-0062
VL - 43
SP - 3608
EP - 3620
JO - IEEE Transactions on Medical Imaging
JF - IEEE Transactions on Medical Imaging
IS - 10
ER -