EchoWorld: Learning Motion-Aware World Models for Echocardiography Probe Guidance

CVPR 25 2025 65.7 method, application

TLDR

EchoWorld uses motion-aware world models for echocardiography probe guidance, reducing errors via pre-trained masked prediction and motion-aware attention.

Reasoning

The paper presents a novel application of world models to probe guidance, leveraging a large real-world dataset and achieving significant error reduction. However, it is domain-specific and lacks comparison to reinforcement learning approaches.

Read-first score

Read-first score 65.7, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 39.

Recency 8%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=True; dataset=False; markers=code,github

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=analysis,evaluation

Topical relevance 42%

55.7

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Field roles

FrontierReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 265.

Keyword Scores

world model

world dynamics prediction

world simulator

generative world model

video world model

interactive world model

model-based reinforcement learning world model

Deep Analysis

Innovations

Motion-aware world modeling framework for echocardiography probe guidance that encodes anatomical knowledge and motion-induced visual dynamics
Pre-training strategy inspired by world modeling principles, predicting masked anatomical regions and simulating visual outcomes of probe adjustments
Motion-aware attention mechanism in fine-tuning stage that integrates historical visual-motion data for precise and adaptive probe guidance

Methodology

EchoWorld employs a two-stage approach: first, a world-modeling pre-training on over one million ultrasound images from more than 200 routine scans, where the model learns to predict masked anatomical regions and simulate visual outcomes of probe adjustments. Then, in the fine-tuning stage, a motion-aware attention mechanism integrates historical visual-motion sequences to enable precise probe guidance. The model is evaluated against existing visual backbones and guidance frameworks using both single-frame and sequential evaluation protocols.

Key Results

EchoWorld significantly reduces guidance errors compared to existing visual backbones and guidance frameworks, excelling in both single-frame and sequential evaluation protocols. Qualitative analysis confirms that the model effectively captures key echocardiographic knowledge.

Resources

Paper Code