Low Complexity Point Tracking of the Myocardium in 2D Echocardiography

2025-03-13

AI Summary

Existing deep learning–based point tracking methods for 2D echocardiography suffer from poor domain adaptability, high computational cost, and difficulty balancing speed and accuracy. To address these limitations, this paper proposes MyoTracker—a lightweight point tracking architecture with only 0.3M parameters. MyoTracker innovatively embeds full temporal context into a low-complexity network to enable single-step, whole-sequence trajectory prediction. Built upon a simplified reconstruction of CoTracker2, it integrates long-term temporal modeling with efficient inference mechanisms, enabling—for the first time—right ventricular (RV) myocardial point tracking and strain quantification. Experiments demonstrate an average trajectory error of 2.00 ± 0.53 mm and an RV free-wall strain bias of −0.3% (95% limits of agreement: [−6.1%, 5.4%]), surpassing inter-observer variability among clinicians. Moreover, MyoTracker achieves a 74× speedup in inference over CoTracker2 and reduces GPU memory consumption by 67–84%.

Technology Category

• Computer Vision: Motion & Tracking
• Machine Learning: Mixture of Experts (MoE)
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Security and Privacy: Tracking, profiling, and countermeasures against them
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Deep learning methods for point tracking are applicable in 2D echocardiography, but do not yet take advantage of domain specifics that enable extremely fast and efficient configurations. We developed MyoTracker, a low-complexity architecture (0.3M parameters) for point tracking in echocardiography. It builds on the CoTracker2 architecture by simplifying its components and extending the temporal context to provide point predictions for the entire sequence in a single step. We applied MyoTracker to the right ventricular (RV) myocardium in RV-focused recordings and compared the results with those of CoTracker2 and EchoTracker, another specialized point tracking architecture for echocardiography. MyoTracker achieved the lowest average point trajectory error at 2.00 $pm$ 0.53 mm. Calculating RV Free Wall Strain (RV FWS) using MyoTracker's point predictions resulted in a -0.3$%$ bias with 95$%$ limits of agreement from -6.1$%$ to 5.4$%$ compared to reference values from commercial software. This range falls within the interobserver variability reported in previous studies. The limits of agreement were wider for both CoTracker2 and EchoTracker, worse than the interobserver variability. At inference, MyoTracker used 67$%$ less GPU memory than CoTracker2 and 84$%$ less than EchoTracker on large sequences (100 frames). MyoTracker was 74 times faster during inference than CoTracker2 and 11 times faster than EchoTracker with our setup. Maintaining the entire sequence in the temporal context was the greatest contributor to MyoTracker's accuracy. Slight additional gains can be made by re-enabling iterative refinement, at the cost of longer processing time.

Problem

• Develops MyoTracker for efficient 2D echocardiography point tracking.
• Compares MyoTracker with CoTracker2 and EchoTracker for accuracy.
• Achieves low GPU memory usage and fast inference speeds.

Innovation

• Low-complexity architecture for echocardiography point tracking
• Extended temporal context for single-step sequence predictions
• Reduced GPU memory usage and faster inference speed