Analysis of small three-dimensional fluorescence spectroscopy dataset using migration learning an example of phenol concentration prediction in wastewater.

Three-dimensional fluorescence spectroscopy has been widely used to detect organic pollutants in water. However, the amount of data required for three-dimensional fluorescence spectroscopy analysis is relatively large, and the time cost of sample collection is high. The amount of data has become an unavoidable limitation of spectral analysis. This study takes the detection of phenol in industrial discharge wastewater as an example and proposes a transfer learning method for small fluorescence spectroscopy datasets. First, fluorescence spectra are generated by splitting them into linear combinations of positively and negatively distributed spectra. Then, based on the idea of transfer learning, the generated fluorescence spectra are used to train a task-specific pre-trained model, which is then transferred to the collected spectral dataset. Experimental results show that the prediction performance of the transfer learning method is improved by 50.08 % compared with that obtained by directly training the model using a small amount of spectral data. In addition, when the spectral data remains unchanged, the accuracy of the model can be improved to a certain extent by increasing the amount of spectral data used for pre-training. The transfer learning method proposed in this study further improves the prediction accuracy when data is limited, and the results of verification in real environments are also satisfactory. It provides a feasible solution to the problem of data limitations in three-dimensional fluorescence spectroscopy.

Fano resonance sensing based on coupling of a sub-wavelength grating and an all-dielectric multilayer film under angle modulation.

Based on the diffraction effect of the waveguide grating, a hybrid structure of an all-dielectric multilayer film and a sub-wavelength dielectric grating is proposed. The discrete state provided in the sub-wavelength waveguide grating couples with the continuous state provided in the Fabry-Perot cavity (F-P cavity) containing the photonic crystal to form a Fano resonance. The relationship between the structural parameter and the incident angle is analyzed by numerical simulation. The result shows that the figure of merit (FOM) value is 533.2RIU-1. This structure provides an effective theoretical basis for the realization of Fano resonance in the all-dielectric sensing structure.

[Determination of Virtual Surgery Mass Point Spring Model Parameters Based on Genetic Algorithms].

Mass point-spring model is one of the commonly used models in virtual surgery. However, its model parameters have no clear physical meaning, and it is hard to set the parameter conveniently. We, therefore, proposed a method based on genetic algorithm to determine the mass-spring model parameters. Computer-aided tomography (CAT) data were used to determine the mass value of the particle, and stiffness and damping coefficient were obtained by genetic algorithm. We used the difference between the reference deformation and virtual deformation as the fitness function to get the approximate optimal solution of the model parameters. Experimental results showed that this method could obtain an approximate optimal solution of spring parameters with lower cost, and could accurately reproduce the effect of the actual deformation model as well.

[Research of Feedback Algorithm and Deformable Model Based on Improved Spring-mass Model].

A new diamond-based variable spring-mass model has been proposed in this study. It can realize the deformation simulation for different organs by changing the length of the springs, spring coefficient and initial angle. The virtual spring joined in the model is used to provide constraint and to avoid hyperelastic phenomenon when excessive force appears. It is also used for the calculation of force feedback in the deformation process. With the deformation force feedback algorithm, we calculated the deformation area of each layer through screening effective particles, and contacted the deformation area with the force. This simplified the force feedback algorithm of traditional spring-particle model. The deformation simulation was realized by the PHANTOM haptic interaction devices based on this model. The experimental results showed that the model had the advantage of simple structure and of being easy to implement. The deformation force feedback algorithm reduces the number of the deformation calculation, improves the real-time deformation and has a more realistic deformation effect.