Deep Learning for Spectral Processing in Microbottle Resonator Humidity Sensors

IEEE Sensors Letters 2026

View at publisher
Authors: Huda Adnan Zain, Khurram Karim Qureshi
DOI: 10.1109/lsens.2026.3670903
Cited by: 0
References: 16

Abstract

This work demonstrates a deep learning–assisted microbottle resonator (MBR) humidity sensor coated with agarose and coupled to a tapered fiber. The convolutional neural network (CNN) processes barcoded spectral data, enabling joint analysis of wavelength shifts, intensity variations, and linewidth changes. The sensor exhibits a CNN classification accuracy of 95%. The results confirm the suitability of deep learning for multidimensional spectral processing in whispering gallery mode (WGM) humidity sensing.

Fields of Study & Subjects

Advanced Fiber Optic Sensors Photonic and Optical Devices Microwave and Dielectric Measurement Techniques

Publication details

Journal / source
IEEE Sensors Letters
Volume
10
Issue
4
Pages
1–4
Publication date
4 March 2026
Language
EN