Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines
Uncertainty estimations through approximate Bayesian inference provide interesting insights to deep neural networks' behavior. In unsupervised learning tasks, where expert labels are unavailable, it becomes ever more important to critique the model through uncertainties. This paper presen...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Open Journal of Engineering in Medicine and Biology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10086579/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832582335860572160 |
|---|---|
| author | Enamundram Naga Karthik Farida Cheriet Catherine Laporte |
| author_facet | Enamundram Naga Karthik Farida Cheriet Catherine Laporte |
| author_sort | Enamundram Naga Karthik |
| collection | DOAJ |
| description | Uncertainty estimations through approximate Bayesian inference provide interesting insights to deep neural networks' behavior. In unsupervised learning tasks, where expert labels are unavailable, it becomes ever more important to critique the model through uncertainties. This paper presents a proof-of-concept for generalizing the aleatoric and epistemic uncertainties in unsupervised MR-CT synthesis of scoliotic spines. A novel adaptation of the cycle-consistency constraint in CycleGAN is proposed such that the model predicts the aleatoric uncertainty maps in addition to the standard volume-to-volume translation between Magnetic Resonance (MR) and Computed Tomography (CT) data. Ablation experiments were performed to understand uncertainty estimation as an implicit regularizer and a measure of the model's confidence. The aleatoric uncertainty helps in distinguishing between the bone and soft-tissue regions in CT and MR data during translation, while the epistemic uncertainty provides interpretable information to the user for downstream tasks. |
| format | Article |
| id | doaj-art-c901ed18c7484069aa7c985c6b0eec43 |
| institution | Kabale University |
| issn | 2644-1276 |
| language | English |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Open Journal of Engineering in Medicine and Biology |
| spelling | doaj-art-c901ed18c7484069aa7c985c6b0eec432025-01-30T00:03:49ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762024-01-01542142710.1109/OJEMB.2023.326296510086579Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic SpinesEnamundram Naga Karthik0https://orcid.org/0000-0003-2940-5514Farida Cheriet1Catherine Laporte2https://orcid.org/0000-0002-1029-006XDepartment of Electrical Engineering, École de technologie supérieure, Montréal, CanadaDepartment of Computer Engineering and Software Engineering, Polytechnique Montréal, Montréal, CanadaDepartment of Electrical Engineering, École de technologie supérieure, Montréal, CanadaUncertainty estimations through approximate Bayesian inference provide interesting insights to deep neural networks' behavior. In unsupervised learning tasks, where expert labels are unavailable, it becomes ever more important to critique the model through uncertainties. This paper presents a proof-of-concept for generalizing the aleatoric and epistemic uncertainties in unsupervised MR-CT synthesis of scoliotic spines. A novel adaptation of the cycle-consistency constraint in CycleGAN is proposed such that the model predicts the aleatoric uncertainty maps in addition to the standard volume-to-volume translation between Magnetic Resonance (MR) and Computed Tomography (CT) data. Ablation experiments were performed to understand uncertainty estimation as an implicit regularizer and a measure of the model's confidence. The aleatoric uncertainty helps in distinguishing between the bone and soft-tissue regions in CT and MR data during translation, while the epistemic uncertainty provides interpretable information to the user for downstream tasks.https://ieeexplore.ieee.org/document/10086579/Bayesian UncertaintyGenerative Adversarial NetworksScoliosisInterpretabilityUnsupervised Learning |
| spellingShingle | Enamundram Naga Karthik Farida Cheriet Catherine Laporte Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines IEEE Open Journal of Engineering in Medicine and Biology Bayesian Uncertainty Generative Adversarial Networks Scoliosis Interpretability Unsupervised Learning |
| title | Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines |
| title_full | Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines |
| title_fullStr | Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines |
| title_full_unstemmed | Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines |
| title_short | Uncertainty Estimation in Unsupervised MR-CT Synthesis of Scoliotic Spines |
| title_sort | uncertainty estimation in unsupervised mr ct synthesis of scoliotic spines |
| topic | Bayesian Uncertainty Generative Adversarial Networks Scoliosis Interpretability Unsupervised Learning |
| url | https://ieeexplore.ieee.org/document/10086579/ |
| work_keys_str_mv | AT enamundramnagakarthik uncertaintyestimationinunsupervisedmrctsynthesisofscolioticspines AT faridacheriet uncertaintyestimationinunsupervisedmrctsynthesisofscolioticspines AT catherinelaporte uncertaintyestimationinunsupervisedmrctsynthesisofscolioticspines |