Rotational\nPhotonic Spin Hall Effect Sensor for Liquid\nChemicals Classification via EfficientNet-V2

Photonic spin Hall effect (PSHE) has been widely used\nfor sensing\ntasks; however, its potential appears to be unexplored for the development\nof a compact yet effective sensor for the classification of liquid\nchemicals. In this study, a liquid identification scheme is demonstrated\nbased on the recently proposed rotational PSHE, where the weak measurement\ntechniques are no longer required for sensing. A liquid crystal device\nis fabricated to experimentally validate the rotational PSHE, which\nprovides unique beam patterns for liquid analytes. The collected beam\npattern images are used to train an EfficientNet-V2a fast\nand efficient deep learning architecturefor classifying the\nliquid chemicals. Two groups of liquids are identified with accuracy\nover 99% in the proposed scheme. Moreover, the performances of several\ndeep learning models are compared, demonstrating the fast training\nspeed and high parameter efficiency of the EfficientNet-V2. The proposed\napproach provides an efficient, accurate, and convenient method for\nrefractive index sensing.