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

Rui Wang¹, Yimu Sun¹, Jingxing Guo¹, Huisi Wu^1,*, Jing Qin²

¹College of Computer Science and Software Engineering, Shenzhen University

²Centre for Smart Health, School of Nursing, The Hong Kong Polytechnic University

2400101058@mails.szu.edu.cn, *Corresponding Author: hswu@szu.edu.cn

Figure 1. An illustration of GDKVM architecture. Linear Key-Value Association defines frame-to-frame causal relations as the state transition matrix. Gated Delta Rule helps in dynamically managing memory. Key-Pixel Feature Fusion fuses the local key feature, the global key feature with the pixel feature.

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.

Challenges in Echocardiography Segmentation

Echocardiography segmentation faces several challenges such as low contrast, speckle noise, and signal dropout.

(a)

(b)

(c)

(d)

(e)

(f)

Figure 2. Illustrative challenges for echocardiography video segmentation: (a) speckle noise, (b) indistinct or blurred contours, and (c-f) the substantial changes in the target's shape and scale throughout the cardiac cycle.

BibTeX

@InProceedings{Wang_ICCV25_GDKVM,
    author    = {Wang, Rui and Sun, Yimu and Guo, Jingxing and Wu, Huisi and Qin, Jing},
    title     = {{GDKVM}: Echocardiography Video Segmentation via Spatiotemporal Key-Value Memory with Gated Delta Rule},
    booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
    month     = {October},
    year      = {2025},
    pages     = {12191-12200}
}

Additional Resources

CVF Paper Poster

To-Do List

Problem Formulation Flowchart
Model Processing Flowchart