arXiv 2506.117772026-01-23cs.CVcs.AIcs.CYcs.LG

通过在线聚类蒸馏实现心电图视频的自监督学习

Self-supervised Learning of Echocardiographic Video Representations via Online Cluster Distillation

Department of Engineering Science, University of Oxford（工程科学系，牛津大学）；Nuffield Department of Women’s and Reproductive Health, University of Oxford（妇女与生殖健康系，牛津大学）；Fundamental AI Lab, University of Technology Nuremberg（基础人工智能实验室，纽伦堡技术大学）；University of Amsterdam（阿姆斯特丹大学）

Divyanshu Mishra, Mohammadreza Salehi, Pramit Saha, Olga Patey, Aris T. Papageorghiou, Yuki M. Asano, J. Alison Noble

备注：: Accepted in NeurIPS 2025

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

DISCOVR通过在线聚类蒸馏实现心脏超声视频的自监督学习，结合时序动态和细粒度空间语义，提升临床任务性能。

摘要：

自监督学习（SSL）在自然图像和视频理解领域取得了重大进展，但在如超声心动图（心脏超声）等领域仍面临挑战，原因包括细微的解剖结构、复杂的时序动态以及当前缺乏领域特定的预训练模型。现有的SSL方法如对比学习、掩码建模和基于聚类的方法在高样本间相似性、对低PSNR输入的敏感性或对破坏临床相关特征的激进增强处理上存在困难。我们提出了DISCOVR（用于跨模态视频表示的图像监督蒸馏），一种用于心脏超声视频表示学习的自监督双分支框架。DISCOVR结合了一个基于聚类的视频编码器，用于建模时序动态，以及一个在线图像编码器，用于提取细粒度的空间语义。这两个分支通过语义聚类蒸馏损失连接，将解剖学知识从演进的图像编码器转移到视频编码器，从而实现具有细粒度语义理解的时序一致的表示。在六个覆盖胎儿、儿童和成人人群的超声心动图数据集上进行评估，DISCOVR在零样本和线性探测设置中优于专门的视频异常检测方法和最先进的视频-SSL基线，实现了优越的分割迁移和在临床相关任务如左室射血分数预测上的强大下游性能。代码可在：https://github.com/mdivyanshu97/DISCOVR上获得。

英文摘要：

Self-supervised learning (SSL) has achieved major advances in natural images and video understanding, but challenges remain in domains like echocardiography (heart ultrasound) due to subtle anatomical structures, complex temporal dynamics, and the current lack of domain-specific pre-trained models. Existing SSL approaches such as contrastive, masked modeling, and clustering-based methods struggle with high intersample similarity, sensitivity to low PSNR inputs common in ultrasound, or aggressive augmentations that distort clinically relevant features. We present DISCOVR (Distilled Image Supervision for Cross Modal Video Representation), a self-supervised dual branch framework for cardiac ultrasound video representation learning. DISCOVR combines a clustering-based video encoder that models temporal dynamics with an online image encoder that extracts fine-grained spatial semantics. These branches are connected through a semantic cluster distillation loss that transfers anatomical knowledge from the evolving image encoder to the video encoder, enabling temporally coherent representations enriched with fine-grained semantic understanding.Evaluated on six echocardiography datasets spanning fetal, pediatric, and adult populations, DISCOVR outperforms both specialized video anomaly detection methods and state-of-the-art video-SSL baselines in zero-shot and linear probing setups,achieving superior segmentation transfer and strong downstream performance on clinically relevant tasks such as LVEF prediction. Code available at: https://github.com/mdivyanshu97/DISCOVR

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。