Accelerating Domain-Aware Electron Microscopy Analysis Using Deep Learning Models with Synthetic Data and Image-Wide Confidence Scoring
This work addresses the challenge of domain-aware feature detection in electron microscopy for researchers, though it is incremental as it builds on existing synthetic data methods.
The paper tackled the problem of scarce and flawed manually labeled datasets in microscopy feature detection by developing a physics-based synthetic data generator, achieving comparable precision (0.86), recall (0.63), F1 scores (0.71), and R2=0.82 to human-labeled models, and used confidence scoring to boost performance by 5-30% with a 25% filtering rate.
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.