# IQ-JEPA: A Joint-Embedding Predictive Architecture with a Hermitian Vision Transformer for Sound Speed and Attenuation Estimation from Ultrasound IQ Data

arXiv: https://arxiv.org/abs/2607.22351 · HTML: https://mlsanity.com/paper/2607.22351/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 29 of its 82 references that are in the map.
- First posted on arXiv 24 Jul 2026.

## Builds on (most cited first, 29 of 29)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [Deep Complex Networks](https://mlsanity.com/paper/1705.09792/) (2017)
- [V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised Learning](https://mlsanity.com/paper/2603.14482/) (2026)
- [InversionNet: A Real-Time and Accurate Full Waveform Inversion with CNNs and continuous CRFs](https://mlsanity.com/paper/1811.07875/) (2018)
- [OpenFWI: Large-Scale Multi-Structural Benchmark Datasets for Seismic Full Waveform Inversion](https://mlsanity.com/paper/2111.02926/) (2021)
- [Complex Transformer: A Framework for Modeling Complex-Valued Sequence](https://mlsanity.com/paper/1910.10202/) (2019)
- [Unsupervised Learning of Full-Waveform Inversion: Connecting CNN and Partial Differential Equation in a Loop](https://mlsanity.com/paper/2110.07584/) (2021)
- [Building Blocks for a Complex-Valued Transformer Architecture](https://mlsanity.com/paper/2306.09827/) (2023)
- [From Pretraining to Privacy: Federated Ultrasound Foundation Model with Self-Supervised Learning](https://mlsanity.com/paper/2411.16380/) (2024)
- [A Deep Learning Framework for Single-Sided Sound Speed Inversion in Medical Ultrasound](https://mlsanity.com/paper/1810.00322/) (2018)
- [SeisLM: a Foundation Model for Seismic Waveforms](https://mlsanity.com/paper/2410.15765/) (2024)
- [Image Reconstruction via Variational Network for Real-Time Hand-Held Sound-Speed Imaging](https://mlsanity.com/paper/1807.07416/) (2018)
- [Auto-Linear Phenomenon in Subsurface Imaging](https://mlsanity.com/paper/2305.13314/) (2023)
- [Training Variational Networks with Multi-Domain Simulations: Speed-of-Sound Image Reconstruction](https://mlsanity.com/paper/2006.14395/) (2020)
- [Phase Aberration Correction for in vivo Ultrasound Localization Microscopy Using a Spatiotemporal Complex-Valued Neural Network](https://mlsanity.com/paper/2209.10650/) (2022)
- [Complex Convolutional Neural Networks for Ultrafast Ultrasound Image Reconstruction from In-Phase/Quadrature Signal](https://mlsanity.com/paper/2009.11536/) (2020)
- [Nonlinear Waveform Inversion for Quantitative Ultrasound](https://mlsanity.com/paper/2205.08461/) (2022)
- [Implicit Neural Representations for Speed-of-Sound Estimation in Ultrasound](https://mlsanity.com/paper/2409.14035/) (2024)
- [Self-supervised Deep Learning for Denoising in Ultrasound Microvascular Imaging](https://mlsanity.com/paper/2507.05451/) (2025)
- [Efficient Complex-Valued Vision Transformers for MRI Classification Directly from k-Space](https://mlsanity.com/paper/2601.18392/) (2026)
- [When do complex-valued neural networks help? A study of representation, geometry, and optimization](https://mlsanity.com/paper/2605.27673/) (2026)

## Abstract (arXiv)

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.