M-FANet: Multi-Feature Attention Convolutional Neural Network for Motor Imagery Decoding.

Motor imagery (MI) decoding methods are pivotal in advancing rehabilitation and motor control research. Effective extraction of spectral-spatial-temporal features is crucial for MI decoding from limited and low signal-to-noise ratio electroencephalogram (EEG) signal samples based on brain-computer interface (BCI). In this paper, we propose a lightweight Multi-Feature Attention Neural Network (M-FANet) for feature extraction and selection of multi-feature data. M-FANet employs several unique attention modules to eliminate redundant information in the frequency domain, enhance local spatial feature extraction and calibrate feature maps. We introduce a training method called Regularized Dropout (R-Drop) to address training-inference inconsistency caused by dropout and improve the model's generalization capability. We conduct extensive experiments on the BCI Competition IV 2a (BCIC-IV-2a) dataset and the 2019 World robot conference contest-BCI Robot Contest MI (WBCIC-MI) dataset. M-FANet achieves superior performance compared to state-of-the-art MI decoding methods, with 79.28% 4-class classification accuracy (kappa: 0.7259) on the BCIC-IV-2a dataset and 77.86% 3-class classification accuracy (kappa: 0.6650) on the WBCIC-MI dataset. The application of multi-feature attention modules and R-Drop in our lightweight model significantly enhances its performance, validated through comprehensive ablation experiments and visualizations.

Inhibition characteristics research of aluminum alloy polishing dust explosion through addition of ultrafine Al(OH)3 inerting agent.

Investigations into the deactivation of explosion sensitivity and reduction of flame propagation for aluminium alloy polishing wastes were carried out by the addition of ultrafine Al(OH)3 inerting agent. Meanwhile, high-purity aluminium powders with similar mean diameters were also used as a comparative study. The explosion propagation characteristics of high-purity aluminium dust and aluminium alloy polishing waste dust under different inerting ratios (ε) were tested and investigated using a standardised Hartmann tester and a developed experimental platform. The results show that the minimum ignition energy of high-purity aluminium powder is between 40 and 45 mJ, and the minimum ignition energy of aluminium alloy polishing waste is between 500 and 550 mJ, which is one order of magnitude higher than that of high-purity aluminium powder. The lower explosion limit concentration of aluminium alloy polishing waste dust is 150 g/m3, which is 53.33% of that of high-purity aluminium powder. According to the analysis of the SEM image, the main reason is that the spherical particles of high-purity aluminium dust have a folded surface and good dispersion. Compared with the smooth fibre surface of aluminium alloy polishing waste dust, the former is easier to contact with air and the contact area is larger. Therefore, in engineering practice, it is not appropriate to use high-purity aluminium dust-related explosion parameters as the basis for the risk assessment of combustion and explosion at aluminium alloy polishing work sites. In addition, as the dust concentration decreases, the combustion intensity of high-purity aluminium dust and aluminium alloy polishing waste dust also decreases, and the flame propagation appears to be a discontinuous phenomenon. The peak flame propagation velocity of aluminium alloy polishing waste is 7.368 m/s at a concentration of 300 g/m3, which is 56.85% of that of high-purity aluminium powder. As the inerting ratio increases, the propagation velocity of the explosion flame slows down. When the inerting ratio reaches 30%, the minimum ignition energy of aluminium alloy polishing waste is inerted to 1 J, and self-sustained flame propagation cannot be formed. The results show that the ultra-fine Al(OH)3 powder has a significant inerting effect and is a realistic possibility in the production of aluminium alloy polishing.

Exploring angular-steering illumination-based eyebox expansion for holographic displays.

Holography represents an enabling technology for next-generation virtual and augmented reality systems. However, it remains challenging to achieve both wide field of view and large eyebox at the same time for holographic near-eye displays, mainly due to the essential étendue limitation of existing hardware. In this work, we present an approach to expanding the eyebox for holographic displays without compromising their underlying field of view. This is achieved by utilizing a compact 2D steering mirror to deliver angular-steering illumination beams onto the spatial light modulator in alignment with the viewer's eye movements. To facilitate the same image for the virtual objects perceived by the viewer when the eye moves, we explore an off-axis computational hologram generation scheme. Two bench-top holographic near-eye display prototypes with the proposed angular-steering scheme are developed, and they successfully showcase an expanded eyebox up to 8 mm × 8 mm for both VR- and AR-modes, as well as the capability of representing multi-depth holographic images.

Water and oil-grease barrier properties of PVA/CNF/MBP/AKD composite coating on paper.

In this paper, three kinds of micro-nano bamboo powder (MBP) and alkyl ketene dimer (AKD) were added to the polyvinyl alcohol/cellulose nanofiber (PVA/CNF) coating to prepare PVA/CNF/MBP coated paper and PVA/CNF/M-MBP/AKD coated paper. The results showed that MBP improved the oleophobicity of PVA/CNF coating, and the grease resistance grade of PVA/CNF/B-MBP and PVA/CNF/M-MBP coated papers reached the highest level, with a kit number of 12. Among the PVA/CNF/MBP coated papers, the PVA/CNF/M-MBP coated paper has the best hydrophobic properties, with the water contact angle and Cobb value of 74° and 21.3 g/m2, respectively. In addition, when the AKD dosage was 0.2% in the PVA/CNF/M-MBP/AKD coating, the kit number of the coated paper was 11, the Cobb value was 15.2 g/m2, the water contact angle was 103°, and the tensile strength was found to increase slightly. Therefore, compared with PVA/CNF coated paper, PVA/CNF/M-MBP/AKD coated paper has good strength and excellent hydrophobic and oleophobic properties.

Driving factors influencing spatiotemporal variation of natural organic chlorine in Shennongjia forest soil.

Studying the geochemical behavior of chlorine is the basis of understanding the chlorine cycle in nature. To explore the spatiotemporal distribution of natural organic chlorine (Clorg), L layer (litter fall), F-H layer (humification zone), topsoil layer (0-20 cm), and deep soil layer (20-40 cm) samples were collected from 18 sampling sites at different altitudes (851-2918 m) in Shennongjia Forest in May, August, and December. Clorg content was analyzed, and the Clorg stocks were calculated. The major factors affecting the distribution of Clorg were explored. The results revealed that the sum of Clorg content in four layers varied from 7.958 to 184.686 mg/kg, and the highest value was observed in August. Clorg accounted for 46%-77% of total chlorine, with the highest mean ratio in soil layer (0-20 cm). Clorg content exhibited the following trend: F-H layer > L layer > topsoil layer (0-20 cm) > deep soil layer (20-40 cm). The seasonal patterns of Clorg in soil layers were different from that in L and F-H layers, which were mainly controlled by the content and humification degree of organic matter. Clorg storage was much higher in soil layers (61-246 kg/ha) than those in F-H layer (1.1-7.1 kg/ha) and in L layer (0.1-0.8 kg/ha) because of the large thickness of the soil layers. Overall, the Clorg content exhibited an increasing trend with altitude, except at an altitude of approximately 1800 m. Clorg content in L and F-H layers varied more obviously with altitude than that in soil layers. When inorganic chlorine (Clin) was not a limiting factor for the chlorination process, Clorg content in L and F-H layers was significantly affected by climate and organic matter controlled by altitude, while Clorg content in soil layers was also mediated by metal ions and pH, and soil particle size. This study could provide a scientific basis for assessing the chlorine cycle in nature.

Diffraction model-driven neural network trained using hybrid domain loss for real-time and high-quality computer-generated holography.

Learning-based computer-generated holography (CGH) has demonstrated great potential in enabling real-time, high-quality holographic displays. However, most existing learning-based algorithms still struggle to produce high-quality holograms, due to the difficulty of convolutional neural networks (CNNs) in learning cross-domain tasks. Here, we present a diffraction model-driven neural network (Res-Holo) using hybrid domain loss for phase-only hologram (POH) generation. Res-Holo utilizes the weights of the pretrained ResNet34 as the initialization during the encoder stage of the initial phase prediction network to extract more generic features and also to help prevent overfitting. Also, frequency domain loss is added to further constrain the information that the spatial domain loss is insensitive. The peak signal-to-noise ratio (PSNR) of the reconstructed image is improved by 6.05 dB using hybrid domain loss compared to using spatial domain loss alone. Simulation results show that the proposed Res-Holo can generate high-fidelity 2 K resolution POHs with an average PSNR of 32.88 dB at 0.014 seconds/frame on the DIV2K validation set. Both monochrome and full-color optical experiments show that the proposed method can effectively improve the quality of reproduced images and suppress image artifacts.

Automatic elimination of phase aberrations in digital holography based on Gaussian 1σ- criterion and histogram segmentation.

We propose a numerical and automatic quadratic phase aberration elimination method in digital holography for phase-contrast imaging. A histogram segmentation method based on Gaussian 1σ-criterion is used to obtain the accurate coefficients of quadratic aberrations using the weighted least-squares algorithm. This method needs no manual intervention for specimen-free zone or prior parameters of optical components. We also propose a maximum-minimum-average-standard deviation (MMASD) metric to quantitatively evaluate the effectiveness of quadratic aberration elimination. Simulation and experimental results are demonstrated to verify the efficacy of our proposed method over the traditional least-squares algorithm.

Investigating learning-empowered hologram generation for holographic displays with ill-tuned hardware.

Existing computational holographic displays often suffer from limited reconstruction image quality mainly due to ill-conditioned optics hardware and hologram generation software. In this Letter, we develop an end-to-end hardware-in-the-loop approach toward high-quality hologram generation for holographic displays. Unlike other hologram generation methods using ideal wave propagation, ours can reduce artifacts introduced by both the light propagation model and the hardware setup, in particular non-uniform illumination. Experimental results reveal that, compared with classical computer-generated hologram algorithm counterparts, better quality of holographic images can be delivered without a strict requirement on both the fine assembly of optical components and the good uniformity of laser sources.

Diffraction model-informed neural network for unsupervised layer-based computer-generated holography.

Learning-based computer-generated holography (CGH) has shown remarkable promise to enable real-time holographic displays. Supervised CGH requires creating a large-scale dataset with target images and corresponding holograms. We propose a diffraction model-informed neural network framework (self-holo) for 3D phase-only hologram generation. Due to the angular spectrum propagation being incorporated into the neural network, the self-holo can be trained in an unsupervised manner without the need of a labeled dataset. Utilizing the various representations of a 3D object and randomly reconstructing the hologram to one layer of a 3D object keeps the complexity of the self-holo independent of the number of depth layers. The self-holo takes amplitude and depth map images as input and synthesizes a 3D hologram or a 2D hologram. We demonstrate 3D reconstructions with a good 3D effect and the generalizability of self-holo in numerical and optical experiments.

A large EEG dataset for studying cross-session variability in motor imagery brain-computer interface.

In building a practical and robust brain-computer interface (BCI), the classification of motor imagery (MI) from electroencephalography (EEG) across multiple days is a long-standing challenge due to the large variability of the EEG signals. We collected a large dataset of MI from 5 different days with 25 subjects, the first open-access dataset to address BCI issues across 5 different days with a large number of subjects. The dataset includes 5 session data from 5 different days (2-3 days apart) for each subject. Each session contains 100 trials of left-hand and right-hand MI. In this report, we provide the benchmarking classification accuracy for three conditions, namely, within-session classification (WS), cross-session classification (CS), and cross-session adaptation (CSA), with subject-specific models. WS achieves an average classification accuracy of up to 68.8%, while CS degrades the accuracy to 53.7% due to the cross-session variability. However, by adaptation, CSA improves the accuracy to 78.9%. We anticipate this new dataset will significantly push further progress in MI BCI research in addressing the cross-session and cross-subject challenge.

Cellulose nanofibers/polyvinyl alcohol blends as an efficient coating to improve the hydrophobic and oleophobic properties of paper.

The effect of cellulose nanofibers (CNFs)/polyvinyl alcohol (PVA) coating on the hydrophobic, oleophobic, and strength properties of paper were investigated. The results showed that the size of bamboo fibers (BFs) decreased significantly and the crystallinity increased significantly after biological enzyme treatment. The average length of CNFs obtained by high pressure homogenization was 2.4 µm, the diameter was 28.7 nm, and the crystallinity was 63.63%. When the coating weight of PVA/CNF was 2.0 g/m2 and the CNF dosage was increased from 0.0% to 3.0%, the paper grease resistance grade was increased from 7 to 9, the Cobb value was decreased from 22.68 ± 0.29 g/m2 to 18.37 ± 0.63 g/m2, the contact angle was increased from 67.82° to 93.56°, and the longitudinal and transverse tensile index were increased from 67.72 ± 0.21 N m/g and 37.63 ± 0.25 N m/g to 68.61 ± 0.55 N m/g and 40.71 ± 0.78 N m/g, respectively. When the CNF dosage was 3.0% and the coating weight of PVA/CNF was 4.0 g/m2, the grease resistance grade of the paper was 12, the Cobb value was 21.80 ± 0.39 g/m2, and the longitudinal and transverse tensile indices were 72.11 ± 0.43 N m/g and 42.58 ± 0.48 N m/g, respectively. In summary, the increase of CNFs can effectively improve the lipophobicity, hydrophobicity and tensile strength of the PVA coated paper.

Core-Shell Structured Magnetic Carboxymethyl Cellulose-Based Hydrogel Nanosorbents for Effective Adsorption of Methylene Blue from Aqueous Solution.

This article reports effective removal of methylene blue (MB) dyes from aqueous solutions using a novel magnetic polymer nanocomposite. The core-shell structured nanosorbents was fabricated via coating Fe3O4 nanoparticles with a layer of hydrogel material, that synthesized by carboxymethyl cellulose cross-linked with poly(acrylic acid-co-acrylamide). Some physico-chemical properties of the nanosorbents were characterized by various testing methods. The nanosorbent could be easily separated from aqueous solutions by an external magnetic field and the mass fraction of outer hydrogel shell was 20.3 wt%. The adsorption performance was investigated as the effects of solution pH, adsorbent content, initial dye concentration, and contact time. The maximum adsorption capacity was obtained at neutral pH of 7 with a sorbent dose of 1.5 g L-1. The experimental data of MB adsorption were fit to Langmuir isotherm model and Pseudo-second-order kinetic model with maximum adsorption of 34.3 mg g-1. XPS technique was applied to study the mechanism of adsorption, electrostatic attraction and physically adsorption may control the adsorption behavior of the composite nanosorbents. In addition, a good reusability of 83.5% MB recovering with adsorption capacity decreasing by 16.5% over five cycles of sorption/desorption was observed.

VC1 catalyses a key step in the biosynthesis of vicine in faba bean.

Faba bean (Vicia faba L.) is a widely adapted and high-yielding legume cultivated for its protein-rich seeds1. However, the seeds accumulate the pyrimidine glucosides vicine and convicine, which can cause haemolytic anaemia (favism) in 400 million genetically predisposed individuals2. Here, we use gene-to-metabolite correlations, gene mapping and genetic complementation to identify VC1 as a key enzyme in vicine and convicine biosynthesis. We demonstrate that VC1 has GTP cyclohydrolase II activity and that the purine GTP is a precursor of both vicine and convicine. Finally, we show that cultivars with low vicine and convicine levels carry an inactivating insertion in the coding sequence of VC1. Our results reveal an unexpected, purine rather than pyrimidine, biosynthetic origin for vicine and convicine and pave the way for the development of faba bean cultivars that are free of these anti-nutrients.

Towards a Switchable AR/VR Near-eye Display with Accommodation-Vergence and Eyeglass Prescription Support.

In this paper, we present our novel design for switchable AR/VR near-eye displays which can help solve the vergence-accommodation-conflict issue. The principal idea is to time-multiplex virtual imagery and real-world imagery and use a tunable lens to adjust focus for the virtual display and the see-through scene separately. With this novel design, prescription eyeglasses for near- and far-sighted users become unnecessary. This is achieved by integrating the wearer's corrective optical prescription into the tunable lens for both virtual display and see-through environment. We built a prototype based on the design, comprised of micro-display, optical systems, a tunable lens, and active shutters. The experimental results confirm that the proposed near-eye display design can switch between AR and VR and can provide correct accommodation for both.

Time-multiplexed multi-view three-dimensional display with projector array and steering screen.

A novel time-multiplexed multi-view three-dimensional (3D) display has been implemented using a projector array to provide the image source and an angular steering-screen module to generate multiple high density horizontal views. The liquid crystal (LC)-based steering screen was specially developed to deflect light beams over a small range and operate in synchronism with the projector array with the use of a customized FPGA driver. The prototype produces vivid color 3D scenes with smooth parallax to multiple viewers. The experimental results verify the proposed multi-projection time-multiplexed multi-view 3D display method that uses a steering screen to produce dense views. Displaying both static and dynamic 3D contents is achieved in our implemented 36-view 3D display prototype. The results of crosstalk measurements are given and analyzed to evaluate the display performance.

Iterative phase-retrieval method for generating stereo array of polarization-controlled focal spots.

This Letter introduces an iterative phase-retrieval method based on the Gerchberg-Saxton (G-S) algorithm for generating any arbitrary 3D pattern in image space, while simultaneously controlling the polarization orientation at each pixel. For proof-of-principle, we generate a stereo focal spot array with distinct polarization orientation for each spot. This method is universal for controlling the output polarization; the only requirement is that the input polarization should be spatially inhomogeneous. This work has the potential to impact coherent imaging techniques and spectroscopy.

Computational method for correcting complex optical distortion based on FOV division.

We propose a computational method for correcting complex optical distortion in off-axis optical systems, such as the optical systems found in head-mounted and head-up displays. The proposed method divides the wide field of view (FOV) into subsections, thereby allowing the distortion to be calculated for each small FOV. Instead of applying the conventional distortion model, the distortion coefficients for each small FOV can be calculated using a simple linear polynomial. In addition, in contrast to the conventional distortion coefficients that refer to the deviation between the real and paraxial image, the distortion coefficients employed by this method directly characterize the relationship between the object and its image. Thus, using the polynomial in the reverse manner repeatedly for each small FOV with the corresponding distortion coefficients, a pixel lookup table is obtained, which can be used to accurately compensate for the distortion in the optical system. This method avoids complicated computations, and there are no requirements for intrinsic or extrinsic parameters. Our experiments verified the effectiveness of the method where the root-mean-square deviation of the projected distorted straight lines was corrected from 23 to 65 pixels to approximately 1 pixel.

A 360-degree floating 3D display based on light field regeneration.

Using light field reconstruction technique, we can display a floating 3D scene in the air, which is 360-degree surrounding viewable with correct occlusion effect. A high-frame-rate color projector and flat light field scanning screen are used in the system to create the light field of real 3D scene in the air above the spinning screen. The principle and display performance of this approach are investigated in this paper. The image synthesis method for all the surrounding viewpoints is analyzed, and the 3D spatial resolution and angular resolution of the common display zone are employed to evaluate display performance. The prototype is achieved and the real 3D color animation image has been presented vividly. The experimental results verified the representability of this method.

Omnidirectional-view three-dimensional display system based on cylindrical selective-diffusing screen.

We utilized a high-frame-rate projector, a rotating mirror, and a cylindrical selective-diffusing screen to present a novel three-dimensional (3D) omnidirectional-view display system without the need for any special viewing aids. The display principle and image size are analyzed, and the common display zone is proposed. The viewing zone for one observation place is also studied. The experimental results verify this method, and a vivid color 3D scene with occlusion and smooth parallax is also demonstrated with the system.

Color three-dimensional display with omnidirectional view based on a light-emitting diode projector.

A three-dimensional (3D) color display with 360 degrees omnidirectional views is developed by use of a high frame rate digital micromirror device projector and three-color light-emitting diodes as the light source. The high-speed projector can project a series of color images on a rotating anisotropic diffusing mirror that controls the reflective light in a designed angular region. Observers around the display scene can watch autostereoscopic views from the ominidirectional 360 deg view around the 3D display. The vivid 3D scene can be obtained with perfect color effect and correct parallax, projection effect, and occlusion. The principle of the omnidirectional view 3D color display is analyzed in detail and experimental results of a 3D color object are presented.