A dual adversarial calibration framework for automatic fetal brain biometry
This paper presents a novel approach to automatic fetal brain biometry motivated by needs in low- and medium-income countries. Specifically, we leverage high-end (HE) ultrasound images to build a biometry solution for low-cost (LC) point-of-care ultrasound images. We propose a novel unsupervised dom...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Conference item |
| Language: | English |
| Published: |
IEEE
2021
|
| _version_ | 1797106875684093952 |
|---|---|
| author | Gao, Y Lee, L Droste, R Craik, R Beriwal, S Papageorghiou, A Noble, A |
| author_facet | Gao, Y Lee, L Droste, R Craik, R Beriwal, S Papageorghiou, A Noble, A |
| author_sort | Gao, Y |
| collection | OXFORD |
| description | This paper presents a novel approach to automatic fetal brain biometry motivated by needs in low- and medium-income countries. Specifically, we leverage high-end (HE) ultrasound images to build a biometry solution for low-cost (LC) point-of-care ultrasound images. We propose a novel unsupervised domain adaptation approach to train deep models to be invariant to significant image distribution shift between the image types. Our proposed method, which employs a Dual Adversarial Calibration (DAC) framework, consists of adversarial pathways which enforce model invariance to; i) adversarial perturbations in the feature space derived from LC images, and ii) appearance domain discrepancy. Our Dual Adversarial Calibration method estimates transcerebellar diameter and head circumference on images from low-cost ultrasound devices with a mean absolute error (MAE) of 2.43mm and 1.65mm, compared with 7.28 mm and 5.65 mm respectively for SOTA. |
| first_indexed | 2024-03-07T07:08:40Z |
| format | Conference item |
| id | oxford-uuid:09129891-a88a-4c10-8858-66bd29791223 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:08:40Z |
| publishDate | 2021 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:09129891-a88a-4c10-8858-66bd297912232022-05-31T11:55:35ZA dual adversarial calibration framework for automatic fetal brain biometryConference itemhttp://purl.org/coar/resource_type/c_5794uuid:09129891-a88a-4c10-8858-66bd29791223EnglishSymplectic ElementsIEEE2021Gao, YLee, LDroste, RCraik, RBeriwal, SPapageorghiou, ANoble, AThis paper presents a novel approach to automatic fetal brain biometry motivated by needs in low- and medium-income countries. Specifically, we leverage high-end (HE) ultrasound images to build a biometry solution for low-cost (LC) point-of-care ultrasound images. We propose a novel unsupervised domain adaptation approach to train deep models to be invariant to significant image distribution shift between the image types. Our proposed method, which employs a Dual Adversarial Calibration (DAC) framework, consists of adversarial pathways which enforce model invariance to; i) adversarial perturbations in the feature space derived from LC images, and ii) appearance domain discrepancy. Our Dual Adversarial Calibration method estimates transcerebellar diameter and head circumference on images from low-cost ultrasound devices with a mean absolute error (MAE) of 2.43mm and 1.65mm, compared with 7.28 mm and 5.65 mm respectively for SOTA. |
| spellingShingle | Gao, Y Lee, L Droste, R Craik, R Beriwal, S Papageorghiou, A Noble, A A dual adversarial calibration framework for automatic fetal brain biometry |
| title | A dual adversarial calibration framework for automatic fetal brain biometry |
| title_full | A dual adversarial calibration framework for automatic fetal brain biometry |
| title_fullStr | A dual adversarial calibration framework for automatic fetal brain biometry |
| title_full_unstemmed | A dual adversarial calibration framework for automatic fetal brain biometry |
| title_short | A dual adversarial calibration framework for automatic fetal brain biometry |
| title_sort | dual adversarial calibration framework for automatic fetal brain biometry |
| work_keys_str_mv | AT gaoy adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT leel adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT droster adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT craikr adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT beriwals adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT papageorghioua adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT noblea adualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT gaoy dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT leel dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT droster dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT craikr dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT beriwals dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT papageorghioua dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry AT noblea dualadversarialcalibrationframeworkforautomaticfetalbrainbiometry |