TY - GEN
T1 - Robust 3D Brain MRI Inpainting with Random Masking Augmentation
AU - Zhang, Juexin
AU - Weng, Ying
AU - CHEN, Ke
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.
PY - 2026
Y1 - 2026
N2 - The ASNR-MICCAI BraTS-Inpainting Challenge was established to mitigate dataset biases that limit deep learning models in the quantitative analysis of brain tumor MRI. This paper details our submission to the 2025 challenge, a novel deep learning framework for synthesizing healthy tissue in 3D scans. The core of our method is a U-Net architecture trained to inpaint synthetically corrupted regions, enhanced with a random masking augmentation strategy to improve generalization. Quantitative evaluation confirmed the efficacy of our approach, yielding an SSIM of 0.873 ± 0.004, a PSNR of 24.996 ± 4.694, and an MSE of 0.005 ± 0.087 on the validation set. On the final online test set, our method achieved an SSIM of 0.919 ± 0.088, a PSNR of 26.932 ± 5.057, and an RMSE of 0.052 ± 0.026. This performance secured first place in the BraTS-Inpainting 2025 challenge and surpassed the winning solutions from the 2023 and 2024 competitions on the official leaderboard.
AB - The ASNR-MICCAI BraTS-Inpainting Challenge was established to mitigate dataset biases that limit deep learning models in the quantitative analysis of brain tumor MRI. This paper details our submission to the 2025 challenge, a novel deep learning framework for synthesizing healthy tissue in 3D scans. The core of our method is a U-Net architecture trained to inpaint synthetically corrupted regions, enhanced with a random masking augmentation strategy to improve generalization. Quantitative evaluation confirmed the efficacy of our approach, yielding an SSIM of 0.873 ± 0.004, a PSNR of 24.996 ± 4.694, and an MSE of 0.005 ± 0.087 on the validation set. On the final online test set, our method achieved an SSIM of 0.919 ± 0.088, a PSNR of 26.932 ± 5.057, and an RMSE of 0.052 ± 0.026. This performance secured first place in the BraTS-Inpainting 2025 challenge and surpassed the winning solutions from the 2023 and 2024 competitions on the official leaderboard.
KW - BraTS 2025
KW - Healthy Tissue Synthesis
KW - Inpainting
KW - MRI
UR - https://www.scopus.com/pages/publications/105037995339
U2 - 10.1007/978-3-032-16370-7_9
DO - 10.1007/978-3-032-16370-7_9
M3 - Conference contribution
AN - SCOPUS:105037995339
SN - 9783032163691
T3 - Lecture Notes in Computer Science
SP - 102
EP - 109
BT - Segmentation, Classification, and Synthesis for Brain Tumors and Traumatic Brain Injuries - MICCAI 2025 Challenges
A2 - Bakas, Spyridon
A2 - Dennis, Emily
A2 - Astaraki, Mehdi
A2 - Baid, Ujjwal
A2 - Conte, Gian Marco
A2 - Foltyn-Dumitru, Martha
A2 - Jiang, Zhifan
A2 - Linguraru, Marius George
A2 - Labella, Dominic
A2 - Metz, Marie-Christin
A2 - Anazodo, Udunna
A2 - de Verdier, Maria Correia
A2 - Kofler, Florian
A2 - Li, Hongwei Bran
A2 - Maleki, Nazanin
PB - Springer Science and Business Media Deutschland GmbH
T2 - Brain TumorS Lighthouse Cluster of Challenges, and the Automated Identification of Moderate-Severe Traumatic Brain Injury Lesions Challenge, BraTS 2025 and AIMS-TBI 2025, held in Conjunction International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2025
Y2 - 23 September 2025 through 27 September 2025
ER -