Evaluating different deep learning models for efficient extraction of Raman signals from CARS spectra.

The nonresonant background (NRB) contribution to the coherent anti-Stokes Raman scattering (CARS) signal distorts the spectral line shapes and thus degrades the chemical information. Hence, finding an effective approach for removing NRB and extracting resonant vibrational signals is a challenging task. In this work, a bidirectional LSTM (Bi-LSTM) neural network is explored for the first time to remove the NRB in the CARS spectra automatically, and the results are compared with those of three DL models reported in the literature, namely, convolutional neural network (CNN), long short-term memory (LSTM) neural network, and very deep convolutional autoencoders (VECTOR). The results of the synthetic test data have shown that the Bi-LSTM model accurately extracts the spectral lines throughout the range. In contrast, the other three models' efficiency deteriorated while predicting the peaks on either end of the spectra, which resulted in a 60 times higher mean square error than that of the Bi-LSTM model. The Pearson correlation analysis demonstrated that Bi-LSTM model performance stands out from the rest, where 94% of the test spectra have correlation coefficients of more than 0.99. Finally, these four models were evaluated on four complex experimental CARS spectra, namely, protein, yeast, DMPC, and ADP, where the Bi-LSTM model has shown superior performance, followed by CNN, VECTOR, and LSTM. This comprehensive study provides a giant leap toward simplifying the analysis of complex CARS spectroscopy and microscopy.

Effect of non-resonant background on the extraction of Raman signals from CARS spectra using deep neural networks.

We report the retrieval of the Raman signal from coherent anti-Stokes Raman scattering (CARS) spectra using a convolutional neural network (CNN) model. Three different types of non-resonant backgrounds (NRBs) were explored to simulate the CARS spectra viz (1) product of two sigmoids following the original SpecNet model, (2) Single Sigmoid, and (3) fourth-order polynomial function. Later, 50 000 CARS spectra were separately synthesized using each NRB type to train the CNN model and, after training, we tested its performance on 300 simulated test spectra. The results have shown that imaginary part extraction capability is superior for the model trained with Polynomial NRB, and the extracted line shapes are in good agreement with the ground truth. Moreover, correlation analysis was carried out to compare the retrieved Raman signals to real ones, and a higher correlation coefficient was obtained for the model trained with the Polynomial NRB (on average, ∼0.95 for 300 test spectra), whereas it was ∼0.89 for the other NRBs. Finally, the predictive capability is evaluated on three complex experimental CARS spectra (DMPC, ADP, and yeast), where the Polynomial NRB model performance is found to stand out from the rest. This approach has a strong potential to simplify the analysis of complex CARS spectroscopy and can be helpful in real-time microscopy imaging applications.

Time-Dependent Intensity Ratio-Based Approach for Estimating the Temperature of Laser Produced Plasma.

Reported here is a rapid and simplified approach for modeling the temporal evolution of the plasma temperature. The use of only two emission lines makes this technique simple, accurate, and fast. Usually, multiple emission lines are required for estimating plasma temperature using Boltzmann/Saha-Boltzmann plots. But, in several cases, either multiple emission lines are not available for every element and/or sufficient lines are not free from self-absorption effect. The proposed method greatly increases the possibility of plasma temperature estimation as it requires only two lines. A brass target was used to generate the plasma, using a conventional single-pulse nanosecond laser of ∼7 ns pulse duration at an excitation wavelength of 532 nm. The initial temperature of plasma and the radiation decay constant were estimated using a proposed intensity ratio model. The results were estimated using various combinations of emission lines, which show an excellent agreement with the values obtained using the previously reported method.

Spatio-temporal characterization of ablative Cu plasma produced by femtosecond filaments.

We present the spatial and temporal characterization of the copper (Cu) plasma produced by the femtosecond laser filaments. The filaments of various lengths and intensities were generated with the aid of three different focusing lenses. Further, the filamentation induced breakdown spectroscopy (FIBS) measurements were carried out for each filament at three different positions along the length of the filament. The filaments were spatially characterized by estimating the plasma temperature and electron density. Our investigation has demonstrated that the centre of the filament is the best to obtain a maximum signal. Both the spectral line intensity and their persistence time are highest for the center of the filament. The enhanced persistence and the scalability of the spectral line intensity tested across different focusing geometries can boost the application of this technique in various fields.

A low-cost LIBS detection system combined with chemometrics for rapid identification of plastic waste.

We present, rapid and efficient identification of ten different types of post-consumer plastics obtained from a local recycling unit by deploying a low cost, compact CCD spectrometer in laser-induced breakdown spectroscopy (LIBS) technique. For this investigation, spectral emissions were collected by an Echelle spectrograph equipped with an intensified charge-coupled device (ES-ICCD) as well as a non-gated Czerny Turner CCD spectrometer (NCT-CCD). The performance is evaluated by interrogating the samples in a single-shot as well as accumulation mode (ten consecutive laser shots). The results from principal component analysis (PCA) have shown excellent discrimination. Further, the artificial neural network (ANN) analysis has demonstrated that individual identification accuracies/rates up to ~99 % can be achieved. The data acquired with ES-ICCD in the accumulation of ten shots have shown average identification accuracies ~97 %. Nevertheless, similar performance is achieved with the NCT-CCD spectrometer even in a single shot acquisition which reduces the overall analysis time by a factor of ~15 times compared to the ES-ICCD. Furthermore, the detector/collection system size, weight, and cost also can be reduced by ~10 times by employing a NCT-CCD spectrometer. The results have the potential in realizing a compact and low-cost LIBS system for the rapid identification of plastics with higher accuracies for the real-time application.