Dense Convolutional Neural Network for Identification of Raman Spectra.

The rapid development of cloud computing and deep learning makes the intelligent modes of applications widespread in various fields. The identification of Raman spectra can be realized in the cloud, due to its powerful computing, abundant spectral databases and advanced algorithms. Thus, it can reduce the dependence on the performance of the terminal instruments. However, the complexity of the detection environment can cause great interferences, which might significantly decrease the identification accuracies of algorithms. In this paper, a deep learning algorithm based on the Dense network has been proposed to satisfy the realization of this vision. The proposed Dense convolutional neural network has a very deep structure of over 40 layers and plenty of parameters to adjust the weight of different wavebands. In the kernel Dense blocks part of the network, it has a feed-forward fashion of connection for each layer to every other layer. It can alleviate the gradient vanishing or explosion problems, strengthen feature propagations, encourage feature reuses and enhance training efficiency. The network's special architecture mitigates noise interferences and ensures precise identification. The Dense network shows more accuracy and robustness compared to other CNN-based algorithms. We set up a database of 1600 Raman spectra consisting of 32 different types of liquid chemicals. They are detected using different postures as examples of interfered Raman spectra. In the 50 repeated training and testing sets, the Dense network can achieve a weighted accuracy of 99.99%. We have also tested the RRUFF database and the Dense network has a good performance. The proposed approach advances cloud-enabled Raman spectra identification, offering improved accuracy and adaptability for diverse identification tasks.

Polarization-modulated broadband achromatic bifunctional metasurface in the visible light.

Achromatic bifunctional metasurface is of great significance in optical path miniaturization among advanced integrated optical systems. However, the reported achromatic metalenses mostly utilize a phase compensate scheme, which uses geometric phase to realize the functionality and uses transmission phase to compensate the chromatic aberration. In the phase compensation scheme, all the modulation freedoms of a nanofin are driven at the same time. This makes most of the broadband achromatic metalenses restricted to realizing single function. Also, the phase compensate scheme is always addressed with circularly polarized (CP) incidence, leading to a limitation in efficiency and optical path miniaturization. Moreover, for a bifunctional or multifunctional achromatic metalens, not all the nanofins will work at the same time. Owing to this, achromatic metalenses using a phase compensate scheme are usually of low focusing efficiencies. To this end, based on the pure transmission phase in the x-/y- axis provided by the birefringent nanofins structure, we proposed an all-dielectric polarization-modulated broadband achromatic bifunctional metalens (BABM) in the visible light. Applying two independent phases on one metalens at the same time, the proposed BABM realizes achromatism in a bifunctional metasurface. Releasing the freedom of nanofin's angular orientation, the proposed BABM breaks the dependence on CP incidence. As an achromatic bifunctional metalens, all the nanofins on the proposed BABM can work at the same time. Simulation results show that the designed BABM is capable of achromatically focusing the incident beam to a single focal spot and an optical vortex (OV) under the illumination of x- and y-polarization, respectively. In the designed waveband 500 nm (green) to 630 nm (red), the focal planes stay unchanged at the sampled wavelengths. Simulation results prove that the proposed metalens not only realized bifunctional achromatically, but also breaks the dependence of CP incidence. The proposed metalens has a numerical aperture of 0.34 and efficiencies of 33.6% and 34.6%. The proposed metalens has advantages of being flexible, single layer, convenient in manufacturing, and optical path miniaturization friendly, and will open a new page in advanced integrated optical systems.

Perfect vortex beam with polarization-rotated functionality based on single-layer geometric-phase metasurface.

Perfect vortex (PV) beam has seen significant advances in fields like particle manipulation, optical tweezers, and particle trapping, due to the fact that its ring radius is independent of the topological charge. Although geometric-phase metasurfaces have been proposed to generate PV beams, it always relies on circularly or elliptically polarized incident light, which hinders the miniaturization of compact optical devices. Here, using orthogonal decomposition of polarization vectors (ODPV), we proposed a geometric-phase metasurface, which breaks the dependence of circular polarization, to generate PV beam. In the design of the metasurface, we introduced PV phase profiles corresponding to the left-handed circularly polarized (LCP) component and the right-handed circularly polarized (RCP) component into the metasurface based on the principle of ODPV. We further determined the rotation angle of each nanostructure of the metasurface by calculating the argument of the composite vector of LCP and RCP in the transmission field. Simulation results show that the proposed geometric-phase metasurface can generate the PV beam upon the illumination of a linearly polarized incident. Moreover, the PV beam with polarization-rotated functionality is achieved by setting the polarization rotation angle. Furthermore, dual PV beams with orthogonal polarization states is realized at the same time by superimposing two sets of phase profiles on a single metasurface. It is also demonstrated that the PV beam parameters, such as ring radius and/or topological charge, can be set on demand in the metasurface design. The proposed metasurface has the exceptional advantage of high fabrication tolerance and is optical path miniaturization friendly, and will open a new avenue in advanced compact and integrated optical systems.

Broadband achromatic longitudinal bifocal metalens in the visible range based on a single nanofin unit cell.

Metasurfaces provide a remarkable platform to manipulate over phase, amplitude, and polarization flexibly and precisely. Bifocal metalens draws great research interest due to its ability of converging wavefronts to different focal positions horizontally and longitudinally. However, as wavelength of incident light changes, chromatic aberration will cause the focal lengths reliance on the incident wavelength, which will affect the performance of metasurface, especially for longitudinal bifocal metalens. In this work, a broadband achromatic longitudinal bifocal metalens (BALBM) based on single nanofin unit cell is demonstrated. Pancharatnam-Berry (PB) phase is used to converge the incident light. Cross commixed sequence distribution (CCSD) is introduced to control the positions of focal points FLand FRwhen left-handed circularly polarized (LCP) and right-handed circularly polarized (RCP) incident. Propagation phase is used to compensate the phase difference caused by chromatic aberration. Simulation results show that in the continuous wavelength range from 500 nm to 700 nm, the focal point shifts caused by chromatic dispersion are reduced 81% for FL and 83% for FR, respectively. The focal length variations are stabilized to 6.21% for FLand 4.8% for FRcomparing with the focal lengths at the initial wavelength 500 nm. The proposed BALBM brings advances to bifocal metasurfaces in versatile application areas including machine vision, optical computed tomography and microimaging.

Deeply-recursive convolutional neural network for Raman spectra identification.

Raman spectroscopy has been widely used in various fields due to its unique and superior properties. For achieving high spectral identification speeds and high accuracy, machine learning methods have found many applications in this area, with convolutional neural network-based methods showing great advantages. In this study, we propose a Raman spectral identification method using a deeply-recursive convolutional neural network (DRCNN). It has a very deep network structure (up to 16 layers) for improving performance without introducing more parameters for recursive layers, which eases the difficulty of training. We also propose a recursive-supervision extension to ease the difficulty of training. By testing several different open-source spectral databases, DRCNN has achieved higher prediction accuracies and better performance in transfer learning compared with other CNN-based methods. Superior identification performance is demonstrated, especially by identification, for characteristically similar and indistinguishable spectra.

Occlusion disparity refinement for stereo matching through the geometric prior-based adaptive label search.

In stereo matching, occlusion disparity refinement is one of the challenges when attempting to improve disparity accuracy. In order to refine the disparity in occluded regions, a geometric prior guided adaptive label search method and sequential disparity filling strategy are proposed. In our method, considering the scene structural correlation between pixels, the geometric prior information such as image patch similarity, matching distance, and disparity constraint is used in the proposed label search energy function and the disparity labels are searched by superpixel matching. Thus, the reliable disparity labels are adaptively searched and propagated for occlusion filling. In order to improve the accuracy in large occluded regions, by using the proposed sequential filling strategy, occluded regions are decomposed into multiple blocks and filled in multiple steps from the periphery; thus, reliable labels are iteratively propagated to the interior of occluded regions without violating the smooth disparity assumption. Experimental results on the Middlebury V3 benchmark show that, compared with other state-of-the-art algorithms, the proposed method achieves better disparity results under multiple criteria. The proposed method can provide better disparity refinement for typical stereo matching algorithms.