使用带合成数据和全图置信度评分的深度学习模型加速领域感知的电子显微镜分析
Accelerating Domain-Aware Electron Microscopy Analysis Using Deep Learning Models with Synthetic Data and Image-Wide Confidence Scoring
AI总结：
针对机器学习显微镜分析依赖稀缺人工标注数据、缺乏领域感知的问题，研究提出基于物理的合成数据生成器，结合全图置信度阈值过滤，实现了与人工标注训练相当的性能，且过滤后性能提升5%-30%。
摘要：
机器学习（ML）模型的集成提升了显微镜图像中特征检测的效率、经济性与可靠性，但这类模型的开发与适用性受限于对稀缺且常存在缺陷的人工标注数据集的依赖，以及缺乏领域感知能力。我们通过创建一个基于物理的合成图像与数据生成器解决了这些挑战，最终得到的机器学习模型在精确率（0.86）、召回率（0.63）、F1分数（0.71）以及工程性能预测（R²=0.82）上，均达到了与经人工标注数据训练的模型相当的水平。我们通过利用特征预测置信度分数推导出全图置信度指标，对两个模型均进行了增强，该指标支持通过简单阈值过滤来排除模糊图像和域外图像，在25%的过滤率下实现了5%-30%的性能提升。我们的研究表明，合成数据可以消除机器学习对人工的依赖，并为需要每张图像进行大量特征检测的场景提供了一种实现领域感知的途径。
英文摘要：
The integration of machine learning (ML) models enhances the efficiency, affordability, and reliability of feature detection in microscopy, yet their development and applicability are hindered by the dependency on scarce and often flawed manually labeled datasets and a lack of domain awareness. We addressed these challenges by creating a physics-based synthetic image and data generator, resulting in a machine learning model that achieves comparable precision (0.86), recall (0.63), F1 scores (0.71), and engineering property predictions (R2=0.82) to a model trained on human-labeled data. We enhanced both models by using feature prediction confidence scores to derive an image-wide confidence metric, enabling simple thresholding to eliminate ambiguous and out-of-domain images resulting in performance boosts of 5-30% with a filtering-out rate of 25%. Our study demonstrates that synthetic data can eliminate human reliance in ML and provides a means for domain awareness in cases where many feature detections per image are needed.
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。
