TY - GEN
T1 - Reversing the Abnormal
T2 - 26th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2023
AU - Bercea, Cosmin I.
AU - Wiestler, Benedikt
AU - Rueckert, Daniel
AU - Schnabel, Julia A.
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.
PY - 2023
Y1 - 2023
N2 - Early and accurate disease detection is crucial for patient management and successful treatment outcomes. However, the automatic identification of anomalies in medical images can be challenging. Conventional methods rely on large labeled datasets which are difficult to obtain. To overcome these limitations, we introduce a novel unsupervised approach, called PHANES (Pseudo Healthy generative networks for ANomaly Segmentation). Our method has the capability of reversing anomalies, i.e., preserving healthy tissue and replacing anomalous regions with pseudo-healthy (PH) reconstructions. Unlike recent diffusion models, our method does not rely on a learned noise distribution nor does it introduce random alterations to the entire image. Instead, we use latent generative networks to create masks around possible anomalies, which are refined using inpainting generative networks. We demonstrate the effectiveness of PHANES in detecting stroke lesions in T1w brain MRI datasets and show significant improvements over state-of-the-art (SOTA) methods. We believe that our proposed framework will open new avenues for interpretable, fast, and accurate anomaly segmentation with the potential to support various clinical-oriented downstream tasks. Code: https://github.com/ci-ber/PHANES
AB - Early and accurate disease detection is crucial for patient management and successful treatment outcomes. However, the automatic identification of anomalies in medical images can be challenging. Conventional methods rely on large labeled datasets which are difficult to obtain. To overcome these limitations, we introduce a novel unsupervised approach, called PHANES (Pseudo Healthy generative networks for ANomaly Segmentation). Our method has the capability of reversing anomalies, i.e., preserving healthy tissue and replacing anomalous regions with pseudo-healthy (PH) reconstructions. Unlike recent diffusion models, our method does not rely on a learned noise distribution nor does it introduce random alterations to the entire image. Instead, we use latent generative networks to create masks around possible anomalies, which are refined using inpainting generative networks. We demonstrate the effectiveness of PHANES in detecting stroke lesions in T1w brain MRI datasets and show significant improvements over state-of-the-art (SOTA) methods. We believe that our proposed framework will open new avenues for interpretable, fast, and accurate anomaly segmentation with the potential to support various clinical-oriented downstream tasks. Code: https://github.com/ci-ber/PHANES
KW - Generative Networks
KW - Unsupervised Anomaly Detection
UR - http://www.scopus.com/inward/record.url?scp=85174691721&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-43904-9_29
DO - 10.1007/978-3-031-43904-9_29
M3 - Conference contribution
AN - SCOPUS:85174691721
SN - 9783031439032
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 293
EP - 303
BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2023 - 26th International Conference, Proceedings
A2 - Greenspan, Hayit
A2 - Greenspan, Hayit
A2 - Madabhushi, Anant
A2 - Mousavi, Parvin
A2 - Salcudean, Septimiu
A2 - Duncan, James
A2 - Syeda-Mahmood, Tanveer
A2 - Taylor, Russell
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 8 October 2023 through 12 October 2023
ER -