arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.21769cs.CVcs.LGeess.IV
收藏

BertsWin: 通过组件平衡的结构优化解决3D掩码自编码器中的拓扑稀疏性

BertsWin: Resolving Topological Sparsity in 3D Masked Autoencoders via Component-Balanced Structural Optimization

Moscow Institute of Physics and Technology(莫斯科物理技术学院)

Evgeny Alves Limarenko, Anastasiia Studenikina

AI总结:

BertsWin通过组件平衡的结构优化提升3D掩码自编码器的拓扑稀疏性处理,实现比标准ViT-MAE快5.8倍的语义收敛,并减少15倍的训练轮数。

摘要:

自监督学习(SSL)和视觉Transformer(ViTs)方法在2D医学影像领域展现出有前景的结果,但这些方法在3D体积影像上的应用却充满困难。标准的掩码自编码器(MAE)是2D领域的最佳解决方案,但在预训练过程中,当75%的token被丢弃时,难以捕捉三维空间关系。我们提出了BertsWin,一种结合全BERT式token掩码和Swin Transformer窗口的混合架构,以增强SSL预训练期间3D中的空间上下文学习。与经典MAE不同,BertsWin引入了一个完整的3Dtoken网格(被掩码和可见的),保留了空间拓扑。为了平滑ViT的二次复杂度,使用了单层局部Swin窗口。我们引入了结构优先损失函数,并评估了锥束计算机断层扫描(CT)的颞下颌关节(TMJ)结果。随后的评估包括在3DCT扫描中对TMJ的分割。我们证明,BertsWin架构通过保持完整的三维空间拓扑,比标准ViT-MAE基线加速了语义收敛约5.8倍。此外,当与我们提出的GradientConductor优化器结合时,完整的BertsWin框架在达到最先进的重建保真度所需训练轮数上实现了15倍的减少(44 vs 660)。分析显示,BertsWin实现这一加速而没有通常与密集体积处理相关的计算惩罚。在典型输入分辨率下,该架构在理论FLOP上与稀疏ViT基线保持一致,从而由于更快的收敛而显著减少了总计算资源。

英文摘要:

The application of self-supervised learning (SSL) and Vision Transformers (ViTs) approaches demonstrates promising results in the field of 2D medical imaging, but the use of these methods on 3D volumetric images is fraught with difficulties. Standard Masked Autoencoders (MAE), which are state-of-the-art solution for 2D, have a hard time capturing three-dimensional spatial relationships, especially when 75% of tokens are discarded during pre-training. We propose BertsWin, a hybrid architecture combining full BERT-style token masking using Swin Transformer windows, to enhance spatial context learning in 3D during SSL pre-training. Unlike the classic MAE, which processes only visible areas, BertsWin introduces a complete 3D grid of tokens (masked and visible), preserving the spatial topology. And to smooth out the quadratic complexity of ViT, single-level local Swin windows are used. We introduce a structural priority loss function and evaluate the results of cone beam computed tomography of the temporomandibular joints. The subsequent assessment includes TMJ segmentation on 3D CT scans. We demonstrate that the BertsWin architecture, by maintaining a complete three-dimensional spatial topology, inherently accelerates semantic convergence by a factor of 5.8x compared to standard ViT-MAE baselines. Furthermore, when coupled with our proposed GradientConductor optimizer, the full BertsWin framework achieves a 15-fold reduction in training epochs (44 vs 660) required to reach state-of-the-art reconstruction fidelity. Analysis reveals that BertsWin achieves this acceleration without the computational penalty typically associated with dense volumetric processing. At canonical input resolutions, the architecture maintains theoretical FLOP parity with sparse ViT baselines, resulting in a significant net reduction in total computational resources due to faster convergence.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑