用谷歌医影大模型检测骨骼X光片异常，效果优于传统方法。
Google-MedGemma Based Abnormality Detection in Musculoskeletal radiographs
- 基于MedGemma视觉编码器提取医学影像特征，轻量分类器判别异常。
- 在MURA数据集上表现超越传统卷积与自编码器模型。
- 适合医疗影像自动化分析、放射科辅助诊断场景。
本文提出一种基于MedGemma的骨骼X光片异常自动检测框架。不同于传统自编码器与神经网络流程，该方法采用在多元医学影像上预训练的SigLIP衍生视觉编码器，将预处理后的X光片编码为高维嵌入，并通过轻量多层感知机完成二分类。实验表明，该方法性能显著优于传统卷积与自编码器模型。同时，利用MedGemma的迁移学习能力，提升模型泛化性并优化特征工程。结合现代医学基础模型，不仅增强表征学习，还可实现模块化训练策略，如选择性解冻编码器块以高效适配新领域。结果表明，基于MedGemma的分类系统可推动临床放射图像分诊，具备可扩展、高精度异常检测能力，未来有望应用于更广泛的医学影像自动化分析。
原文摘要 · Abstract (English)
This paper proposes a MedGemma-based framework for automatic abnormality detection in musculoskeletal radiographs. Departing from conventional autoencoder and neural network pipelines, the proposed method leverages the MedGemma foundation model, incorporating a SigLIP-derived vision encoder pretrained on diverse medical imaging modalities. Preprocessed X-ray images are encoded into high-dimensional embeddings using the MedGemma vision backbone, which are subsequently passed through a lightweight multilayer perceptron for binary classification. Experimental assessment reveals that the MedGemma-driven classifier exhibits strong performance, exceeding conventional convolutional and autoencoder-based metrics. Additionally, the model leverages MedGemma's transfer learning capabilities, enhancing generalization and optimizing feature engineering. The integration of a modern medical foundation model not only enhances representation learning but also facilitates modular training strategies such as selective encoder block unfreezing for efficient domain adaptation. The findings suggest that MedGemma-powered classification systems can advance clinical radiograph triage by providing scalable and accurate abnormality detection, with potential for broader applications in automated medical image analysis. Keywords: Google MedGemma, MURA, Medical Image, Classification.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。