Hybrid Data Augmentation Method for Multimode Sensing in a Whispering Gallery Mode Resonator

Abstract

Multimode sensing based on whispering gallery mode (WGM) microcavities has attracted significant attention due to its diverse optical modes, which offer the potential for enhanced sensing performance. While AI models excel at integrating multimodal information, their effectiveness in multimode detection is often limited by the scarcity of large-scale, high-quality training data. To address this issue, we propose a hybrid data augmentation (DA) method that leverages an autoencoder (AE) to generate high-quality synthetic spectral data. Small Gaussian noise is then added to both real experimental and synthetic spectra, creating a large number of optical spectra sets. This significantly expands the training set, thereby improving the regression model’s accuracy and reducing overfitting. When applied to a multimode temperature regression task using a microbubble resonator (MBR), the hybrid DA method results in a 85.91% improvement in the prediction accuracy of the deep neural network (DNN). This approach is simple, efficient, and well-suited for multi-parameter regression tasks.