无需训练，用DINOv3在测试时优化配准，医学影像对齐更精准。
DINOv3 with Test-Time Training for Medical Image Registration
- 冻结DINOv3编码器，测试时直接优化特征空间的形变场。
- 腹部MR-CT数据上达0.790的Dice分数，HD95仅4.9±5.0。
- 适合临床部署，无须额外标注数据，通用性强。
以往基于学习的医学图像配准方法通常需要大量训练数据，限制了临床应用。为此，我们提出一种无需训练的流程，依赖冻结的DINOv3编码器，并在测试时对特征空间中的形变场进行优化。在两个代表性基准上，该方法表现准确且形变规则。在Abdomen MR-CT数据集上，平均Dice分数（DSC）达到0.790，95%分位豪斯多夫距离（HD95）为4.9±5.0，对数雅可比行列式标准差（SDLogJ）为0.08±0.02，均为最优。在ACDC心脏MRI数据集上，平均DSC提升至0.769，SDLogJ降至0.11，HD95降至4.8，显著优于初始对齐效果。结果表明，在测试时于紧凑的基础特征空间操作，是一种无需额外训练、具实际意义且通用的临床配准解决方案。
原文摘要 · Abstract (English)
Prior medical image registration approaches, particularly learning-based methods, often require large amounts of training data, which constrains clinical adoption. To overcome this limitation, we propose a training-free pipeline that relies on a frozen DINOv3 encoder and test-time optimization of the deformation field in feature space. Across two representative benchmarks, the method is accurate and yields regular deformations. On Abdomen MR-CT, it attained the best mean Dice score (DSC) of 0.790 together with the lowest 95th percentile Hausdorff Distance (HD95) of 4.9+-5.0 and the lowest standard deviation of Log-Jacobian (SDLogJ) of 0.08+-0.02. On ACDC cardiac MRI, it improves mean DSC to 0.769 and reduces SDLogJ to 0.11 and HD95 to 4.8, a marked gain over the initial alignment. The results indicate that operating in a compact foundation feature space at test time offers a practical and general solution for clinical registration without additional training.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。