NVAM-Net: deep learning networks for reconstructing high-quality fiber orientation distributions.

PURPOSE: Diffusion magnetic resonance imaging (dMRI) is a widely used non-invasive method for investigating brain anatomical structures. Conventional techniques for estimating fiber orientation distribution (FOD) from dMRI data often neglect voxel-level spatial relationships, leading to ambiguous associations between target voxels and their neighbors, which, in turn, adversely impacts FOD accuracy. This study aims to address this issue by introducing a novel neural network, the neighboring voxel attention mechanism network (NVAM-Net), designed to reconstruct high-quality FOD images. METHODS: The NVAM-Net leverages a Transformer architecture and incorporates two innovative attention mechanisms: voxel attention and surface attention. These mechanisms are specifically designed to capture overlooked features among neighboring voxels. The processed features are subsequently passed through two fully connected layers, further enhancing FOD estimation accuracy by separately estimating spherical harmonics (SH) coefficients of varying orders. RESULTS: The experimental findings, based on the Human Connectome Project (HCP) dataset, reveal that the reconstructed super-resolution FOD images achieve results comparable to those obtained through more advanced dMRI acquisition protocols. These results underscore the NVAM-Net's robust performance in reconstructing multi-shell multi-tissue constrained spherical deconvolution (MSMT-CSD). CONCLUSION: In summary, this research underscores the NVAM-Net's advantages and practical feasibility in reconstructing high-quality FOD images. It provides a reliable reference point for clinical applications in the field of diffusion magnetic resonance imaging.

Enhanced photovoltaic performance of nanowire array solar cells with multiple diameters.

A multi-diameter p-i-n junction GaAs nanowire (NW) array architecture is proposed for high-performance solar cells. Coupled three-dimensional optoelectronic simulations are performed to investigate the photovoltaic properties. The NW diameters are randomly selected within the range of 220-400 nm, following the Gaussian distribution. The results show that the absorption strongly depends on the diameter, and the multi-diameter NW array exhibits higher optical absorption, in comparison with the uniform-diameter counterpart. This is because of the superposition of multiple absorption peaks. Moreover, the multi-diameter NW array can efficiently enhance the effective absorption; that is, the depletion region absorption, which directly leads to increased photocurrent. A remarkable efficiency of 17% is obtained for a 16-diameter NW array solar cell with a full width at half maximum of the diameter distribution of 75 nm, higher than the best value (16.1%) of uniform-diameter device with an optimum diameter of 310 nm. This work demonstrates that the native diameter nonuniformity of self-organized nanowires is beneficial for high-performance photovoltaics with low cost and a simple fabrication process.

MFCA-MICNN: a convolutional neural network with multiscale fast channel attention and multibranch irregular convolution for noise removal in dMRI.

Diffusion magnetic resonance imaging (dMRI) currently stands as the foremost noninvasive method for quantifying brain tissue microstructure and reconstructing white matter fiber pathways. However, the inherent free diffusion motion of water molecules in dMRI results in signal decay, diminishing the signal-to-noise ratio (SNR) and adversely affecting the accuracy and precision of microstructural data. In response to this challenge, we propose a novel method known as the Multiscale Fast Attention-Multibranch Irregular Convolutional Neural Network for dMRI image denoising. In this work, we introduce Multiscale Fast Channel Attention, a novel approach for efficient multiscale feature extraction with attention weight computation across feature channels. This enhances the model's capability to capture complex features and improves overall performance. Furthermore, we propose a multi-branch irregular convolutional architecture that effectively disrupts spatial noise correlation and captures noise features, thereby further enhancing the denoising performance of the model. Lastly, we design a novel loss function, which ensures excellent performance in both edge and flat regions. Experimental results demonstrate that the proposed method outperforms other state-of-the-art deep learning denoising methods in both quantitative and qualitative aspects for dMRI image denoising with fewer parameters and faster operational speed.

[Analysis of rhythm features of EEG for driving fatigue].

With extracting separately delta, theta, alpha and beta rhythms of electroencephalogram (EEG), we studied the characters of EEG for fatigued drivers by analyzing relative power spectrum, power spectral entropy and brain electrical activity mapping. The experimental results showed that with the average relative power spectrum in delta and theta rhythms of EEG increasing, the average relative power spectrum in alpha and beta rhythms decreased, while the average relative power spectrum in delta, theta and alpha rhythms increased in deep fatigue. The average power spectral entropy of EEG decreases with the increasing fatigue level. The average relative power spectrum and the average power spectral entropy of EEG could be expected to serve as the index for detecting fatigue level of drivers.

ACTNet: asymmetric convolutional transformer network for diabetic foot ulcers classification.

Most existing image classification methods have achieved significant progress in the field of natural images. However, in the field of diabetic foot ulcer (DFU) where data is scarce and complex, the accurate classification of data is still a thorny problem. In this paper, we propose an Asymmetric Convolutional Transformer Network (ACTNet) for the multi-class (4-class) classification task of DFU. Specifically, in order to strengthen the expressive ability of the network, we design an asymmetric convolutional module in the front part of the network to model the relationship between local pixels, extract the underlying features of the image, and guide the network to focus on the central region in the image that contains more information. Furthermore, a novel pooling layer is added between the encoder and the classification head in the Transformer, which weights the data sequence generated by the encoder to better correlate the features between the input data. Finally, to fully exploit the performance of the model, we pretrained our model on ImageNet and fine-tune it on DFU images. The model is validated on the DFUC2021 test set, and the F1-score and AUC value are 0.593 and 0.824, respectively. The experiments show that our model has excellent performance even in the case of a small dataset.

[Application of HHT to driving fatigue in EEG analysis].

Based on the fact that the signals of electroencephalogram (EEG) possess non-linear and non-stationary properties, Hilbert-Huang Transform (HHT) was proposed for the EEG analysis of driving fatigue. Firstly, C4-lead EEG was selected, and the data of normal driving state and fatigue driving state was analyzed by HHT to explore the differences. Then O2-lead EEG was chosen for contrastive analysis of differences between the different leads. It was found through the analysis that the EEG signals had different Hilbert marginal spectrums for different states, and there were also some differences at the same state for the two leads. It can be certain that HHT can well distinguish different states of drivers as a novel approach for driving fatigue detection, and the selected lead may affect detectable results to some extent.

[A new method of tremor diagnosis based on singular value decomposition of EMD].

Aiming at three kinds of tremor, including essential tremor (ET), Parkinsonian disease (PD) tremor and physiological tremor (PT), which are subjected to frequent clinical misdiagnosis, a new method based on singular value decomposition (SVD) of empirical mode decomposition (EMD) and support vector machine (SVM) for the recognition of tremor is proposed in this paper. First, the hand acceleration signals of three different types of 40 tremor voluntary subjects were collected on the basis of informed consent, and the EMD method was used to decompose the signals into a number of stationary intrinsic mode functions (IMFs). Then the preceding four IMFs which could describe signals were selected, and the initial feature vector matrixes were formed. After the application of SVD technique to the initial feature vector matrixes, the singular values were obtained and used as the feature vectors of tremor types to be put in the support vector machine classifier as well as in the identification of tremor types. The results of experiment have shown that the proposed diagnosis method based on SVD of EMD and SVM can extract tremor features effectively and identify tremor types accurately. It also provides a new assistant approach for clinical diagnosis of tremor.

Photovoltaic Performance of Pin Junction Nanocone Array Solar Cells with Enhanced Effective Optical Absorption.

The photovoltaic performance of axial and radial pin junction GaAs nanocone array solar cells is investigated. Compared with the cylinder nanowire arrays, the nanocone arrays not only improve the whole optical absorption but more importantly enhance the effective absorption (absorption in the depletion region). The enhanced effective absorption is attributed to the downward shift and extension of the absorption region induced by the shrinking top, which dramatically suppresses the absorption loss in the high-doped top region and enhances the absorption in the depletion region. The highest conversion efficiencies for axial and radial GaAs nanocone solar cells are 20.1% and 17.4%, obtained at a slope angle of 5° and 6°, respectively, both of which are much higher than their cylinder nanowire counterparts. The nanocone structures are promising candidates for high-efficiency solar cells.

[The current situation and development of tremor signals analysis].

Tremor, a rhythmic and involuntary oscillatory movement of one or several body parts, is the movement resulting from the abnormal synchronization of motor neural units. Detecting and analyzing the ACC, EMG and EEG signals of tremor patients by signal processing methods are very important for clinical diagnosis, rating evaluation and detection of incipient illness. This paper introduces the applications of time domain,frequency domain, artificial neural network, high order accumulation, approximate entropy, fuzzy, chaos, discriminant analysis in the researches of tremor signals, and finally points out the application foreground of researches on tremor signals.