GDKVM: Echocardiography Video Segmentation via Spatiotemporal Key-Value Memory with Gated Delta Rule

2025-12-10

AI Summary

Cardiac ultrasound video segmentation of the left ventricular cavity is challenged by motion artifacts, noise, and myocardial deformation, leading to temporal inconsistency and ambiguous boundaries. To address these issues, we propose LKV-Net: (1) a novel Linear Key-Value mechanism to model long-range temporal dependencies across frames; (2) a Gated Dynamic Refresh (GDR) rule for efficient memory state management; (3) a Multi-Scale Key Pixel Feature Fusion (KPFF) module to enhance robustness against artifacts and variability; and (4) lightweight spatiotemporal modeling for efficient, fine-grained segmentation. Evaluated on CAMUS and EchoNet-Dynamic benchmarks, LKV-Net achieves state-of-the-art performance—improving Dice score by 3.2%, significantly enhancing noise robustness, and maintaining real-time inference speed. The method thus advances both clinical accuracy and practical deployability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements

Abstract

Accurate segmentation of cardiac chambers in echocardiography sequences is crucial for the quantitative analysis of cardiac function, aiding in clinical diagnosis and treatment. The imaging noise, artifacts, and the deformation and motion of the heart pose challenges to segmentation algorithms. While existing methods based on convolutional neural networks, Transformers, and space-time memory networks have improved segmentation accuracy, they often struggle with the trade-off between capturing long-range spatiotemporal dependencies and maintaining computational efficiency with fine-grained feature representation. In this paper, we introduce GDKVM, a novel architecture for echocardiography video segmentation. The model employs Linear Key-Value Association (LKVA) to effectively model inter-frame correlations, and introduces Gated Delta Rule (GDR) to efficiently store intermediate memory states. Key-Pixel Feature Fusion (KPFF) module is designed to integrate local and global features at multiple scales, enhancing robustness against boundary blurring and noise interference. We validated GDKVM on two mainstream echocardiography video datasets (CAMUS and EchoNet-Dynamic) and compared it with various state-of-the-art methods. Experimental results show that GDKVM outperforms existing approaches in terms of segmentation accuracy and robustness, while ensuring real-time performance. Code is available at https://github.com/wangrui2025/GDKVM.

Problem

• Segments cardiac chambers in echocardiography videos accurately
• Addresses noise, artifacts, and heart motion challenges in segmentation
• Balances spatiotemporal dependency capture with computational efficiency

Innovation

• Linear Key-Value Association models inter-frame correlations
• Gated Delta Rule efficiently stores intermediate memory states
• Key-Pixel Feature Fusion integrates multi-scale local and global features