IQ-JEPA: A Joint-Embedding Predictive Architecture with a Hermitian Vision Transformer for Sound Speed and Attenuation Estimation from Ultrasound IQ Data
作者: Masashi Sode, Gianmarco Pinton
分类: cs.LG, physics.med-ph
发布日期: 2026-07-24
💡 一句话要点
提出IQ-JEPA以解决超声IQ数据中的声速和衰减估计问题
🎯 匹配领域: 支柱二：RL算法与架构 (RL & Architecture) 支柱八：物理动画 (Physics-based Animation) 支柱九：具身大模型 (Embodied Foundation Models)
关键词: 超声成像 声速估计 自监督学习 Hermitian视觉变换器 标签效率 定量超声 机器学习
📋 核心要点
- 声速的估计是超声成像中的一大挑战，现有方法对标签的需求高，且模拟标签成本昂贵。
- IQ-JEPA通过无标签预训练和有标签微调的方式，利用了丰富的真实通道数据和少量的模拟标签。
- 在79,293次Fullwave 2.5模拟实验中，该方法在10,000个标签下达到了15.60 m/s的声速估计，标签效率提升约三倍。
📝 摘要（中文）
声速在组织成像中至关重要，但从原始脉冲回波通道数据中恢复声速是一个非线性逆问题。现有的学习求解器速度快但对标签需求高，而模拟声速标签昂贵，真实通道数据则缺乏标签。本文提出IQ-JEPA，通过无标签预训练和模拟图的微调，利用了这两种数据类型。该方法使用Hermitian视觉变换器直接处理复杂信号，显著提高了标签效率，并在大量实验中展示了自监督学习的优势。
🔬 方法详解
问题定义：本文旨在解决从超声IQ数据中估计声速和衰减的问题。现有方法通常需要大量的标签数据，而模拟标签的获取成本高，导致实际应用受限。
核心思路：IQ-JEPA通过无标签预训练来预测被遮蔽的IQ区域的潜在表示，然后在模拟声速图上进行微调。这种方法有效利用了丰富的无标签数据，降低了对标签的依赖。
技术框架：该方法的整体架构包括两个主要阶段：首先是无标签的预训练阶段，使用Hermitian视觉变换器直接处理复杂信号；其次是有标签的微调阶段，针对模拟声速图进行优化。
关键创新：最重要的创新在于使用Hermitian视觉变换器，该变换器的注意力机制对相位差不变，能够直接从IQ信号中提取声速信息，与传统方法相比具有更高的效率和准确性。
关键设计：在网络结构上，Hermitian视觉变换器的设计使其能够处理复杂信号，且其前馈网络对相位的共轭乘积不变。此外，损失函数的设计也考虑了声速和衰减的特征，以提高模型的学习效果。
🖼️ 关键图片
📊 实验亮点
在79,293次Fullwave 2.5模拟实验中，IQ-JEPA在10,000个标签下达到了15.60 m/s的声速估计，标签效率提升约三倍，且在1,000个标签时提升幅度超过四倍。与InversionNet基线相比，性能提升约2.2倍，充分展示了自监督学习的优势。
🎯 应用场景
该研究在医学成像领域具有广泛的应用潜力，尤其是在超声成像中。通过提高声速和衰减的估计精度，可以显著改善成像质量，进而提升诊断的准确性。未来，该方法有望成为定量超声的基础模型，推动相关技术的发展。
📄 摘要（原文）
The speed of sound in tissue is a prerequisite for well-focused imaging and has diagnostic value, but recovering it from raw pulse-echo channel data is fundamentally a nonlinear inverse problem. Learned solvers are fast yet label hungry. Simulated sound-speed labels are expensive, while abundant real channel data is unlabeled. We propose IQ-JEPA to exploit both data types. An encoder is pretrained without labels to predict the latent representation of masked in-phase and quadrature (IQ) regions from visible context, then fine-tuned on simulated maps. Sound speed appears in the IQ signal as a phase difference, invariant to the constant phase offset. The encoder is a Hermitian vision transformer that operates on the complex signal directly. Its attention is equivariant to that phase and its conjugate-product feed-forward is invariant to it, so the encoder reads a quantity analogous to the one classical coherence methods use. On 79,293 Fullwave 2.5 simulations at 2.5 MHz, pretraining on the 63,435 unlabeled acquisitions reaches 15.60 m/s at 10,000 labels. This is a roughly threefold gain in label efficiency over supervised training, growing to over fourfold at 1,000 labels. It is about 2.2x below an InversionNet baseline, and 8.71 m/s at full labels. The gain still grows with more unlabeled pretraining data. Our comparisons point to self-supervision as the dominant factor. The same encoder transfers. Its frozen features expose sound speed and attenuation, and cross-distribution pretraining between layered and abdominal phantoms costs little accuracy. We see this as a first step toward a foundation model for quantitative ultrasound.