[论文解读] Self-Consistent Recursive Diffusion Bridge for Medical Image Translation

Fuat Arslan, Bilal Kabaş+3 more|arXiv (Cornell University)|May 10, 2024

Brain Tumor Detection and Classification被引用 13

一句话总结

SelfRDB 引入一种扩散桥框架，其前向过程以加入噪声的源作为端点，并采用自一致的递归反向过程，从而在跨模态医学影像翻译（MRI-CT，多对比度 MRI）方面取得优越表现。

ABSTRACT

Denoising diffusion models (DDM) have gained recent traction in medical image translation given improved training stability over adversarial models. DDMs learn a multi-step denoising transformation to progressively map random Gaussian-noise images onto target-modality images, while receiving stationary guidance from source-modality images. As this denoising transformation diverges significantly from the task-relevant source-to-target transformation, DDMs can suffer from weak source-modality guidance. Here, we propose a novel self-consistent recursive diffusion bridge (SelfRDB) for improved performance in medical image translation. Unlike DDMs, SelfRDB employs a novel forward process with start- and end-points defined based on target and source images, respectively. Intermediate image samples across the process are expressed via a normal distribution with mean taken as a convex combination of start-end points, and variance from additive noise. Unlike regular diffusion bridges that prescribe zero variance at start-end points and high variance at mid-point of the process, we propose a novel noise scheduling with monotonically increasing variance towards the end-point in order to boost generalization performance and facilitate information transfer between the two modalities. To further enhance sampling accuracy in each reverse step, we propose a novel sampling procedure where the network recursively generates a transient-estimate of the target image until convergence onto a self-consistent solution. Comprehensive analyses in multi-contrast MRI and MRI-CT translation indicate that SelfRDB offers superior performance against competing methods.

研究动机与目标

• 推动源模态与目标模态之间的医学影像翻译超越标准扩散模型的能力。
• 提出一个带有软先验端点的扩散桥，以增强泛化和信息传递。
• 引入一个自一致的递归采样机制，以提升反向步骤的精度。

提出的方法

• 带有目标源凸组合的前向过程：x_t ~ N(mu_x0,t x0 + mu_y,t y, sigma_t^2 I) 且 sigma_t^2 单调增加直至端点（加入噪声的源）。
• 端点为一个加入噪声的源图像 x_T = y_epsilon ，而不是一个硬源样本。
• 反向过程使用恢复网络 G_theta 迭代地产生自一致的目标估计 x0~*，并从后验 q(x_{t-1}|x_t,y,x0~*) 中采样 x_{t-1}。
• 自一致递归：x0~{r+1} = G_theta(x_t,t,y,x0~{r}) 直到收敛到 x0~*；随后在 x0~* 下从 q 中采样 x_{t-1}。
• 对抗判别器 D_theta 引导反向步骤样本的真实感，像素级 L1 损失将恢复的 x0~* 与真值 x0 对齐。

实验结果

研究问题

• RQ1 带有软先验端点的扩散桥是否能够改善源图像与目标图像之间的跨模态翻译？
• RQ2 自一致递归采样方案是否提升多模态翻译中中间样本的准确性和真实感？
• RQ3 相较于GANs和基于扩散的基线，SelfRDB 在多对比度 MRI 以及 MRI-CT 翻译中的表现如何？
• RQ4 端点噪声水平和前向方差调度对对未见源图像的泛化有何影响？

主要发现

• SelfRDB 在多对比度 MRI 和 MRI-CT 任务中超越对比方法（GANs 和扩散模型）。
• 单调增加的端点噪声调度通过软化源模态上的硬先验来提升泛化。
• 自一致递归采样过程提供更准确的反向步估计和更好的图像合成。
• 在 IXI、BRATS 与盆腔 MRI-CT 数据集上，SelfRDB 的 PSNR 和 SSIM 高于 SynDiff、DDPM、I2SB 以及 pix2pix 基线。