Search

Abstract Lung segmentation is an essential step in a computer‐aided diagnosis system for chest radiographs. The lung parenchyma is first segmented in pulmonary computer‐aided diagnosis systems to remove the interference of non‐lung regions while increasing the effectiveness of the subsequent work. Nevertheless, most medical image segmentation methods nowadays use U‐Net and its variants. These variant networks perform poorly in segmentation to detect smaller structures and cannot accurately segment boundary regions. A multi‐interaction feature fusion network model based on Kiu‐Net is presented in this paper to address this problem. Specifically, U‐Net and Ki‐Net are first utilized to extract high‐level and detailed features of chest images, respectively. Then, cross‐residual fusion modules are employed in the network encoding stage to obtain complementary features from these two networks. Second, the global information module is introduced to guarantee the segmented region's integrity. Finally, in the network decoding stage, the multi‐interaction module is presented, which allows to interact with multiple kinds of information, such as global contextual information, branching features, and fused features, to obtain more practical information. The performance of the proposed model was assessed on both the Montgomery County (MC) and Shenzhen datasets, demonstrating its superiority over existing methods according to the experimental results.

Water level prediction is an essential factor for the safe operation of pumping stations. However, due to the complex nonlinear relationship between the water level of the front pool of the pumping station and the influencing factors, the prediction of the water level is still inaccurate and untimely. Backpropagation (BP) neural network, improved particle swarm optimization-BP neural network (PSO-BP), support vector machine regression (SVR), and improved PSO-SVR were used to construct 4-h and 8-h ahead prediction models for pumping station prestation water levels. Mean absolute error, mean squared error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), and R correlation coefficient were used as prediction evaluation metrics. This method is applied in the Baliwan Pumping Station, the highest pumping station in the South-to-North Water Diversion Eastern Route Project (SNWDERP). The results showed that the MSE, RMSE, and MAPE of the improved PSO-BP model were smaller than other models, whereas the R correlation coefficient was larger, confirming its high prediction accuracy. All models had higher prediction accuracy 4 h ahead than 8 h ahead. Combining the time-phased water level prediction method and hybrid machine learning can enhance the water level prediction accuracy of a pumping station pre-station. HIGHLIGHTS We propose a pumping station pre-station water level prediction model based on multiple machine learning algorithms.; The paper categorizes the data into the ice formation period, the aquatic grass period, and the normal scheduling period and compares them to ALL.; The effect of different foresight periods on forecast accuracy was analyzed.;

Abstract Facial expression recognition technology has become a powerful tool for conveying human emotions and intentions and is widely used in areas such as assisted driving and intelligent medical care. Due to the limited computing power of current hardware devices and the real‐time requirements of application scenarios, this paper proposes a high‐performance and lightweight framework for real‐time facial expression recognition framework to solve the problem of real‐time completion of expression recognition tasks under low hardware costs. To address these issues, this paper first designs a RepVGG and mobileNetV2 dual‐channel structure in the feature extraction. It is then input into the MobileViT Block for global feature modelling. Finally, the position vector of the capsule network is used to replace the output of the global pooling, preserving the spatial relationship of the salient features and enhancing the classification effect. Compared with the mainstream facial expression recognition algorithm that cannot get good classification results under low complexity conditions, the model has a significant accuracy improvement while ensuring lightweight. With only 294.60M FLOPS and 0.95M parameters, it achieved an accuracy of 97.53% on the KDEF dataset and 85.56% on the RAF‐DB, demonstrating the advanced nature of the algorithm.

Ear recognition has made good progress as an emerging biometric technology. However, the recognition performance, generalization ability, and feature robustness of ear recognition systems based on hand-crafted features are relatively poor. With the development of deep learning, these problems have been partly overcome. However, the recognition performance of existing ear recognition systems still needs to be improved when facing unconstrained ear databases in realistic scenarios. Another critical problem is that most systems with ear feature template databases are vulnerable to software attacks that disclose users’ privacy and even bring down the system. This paper proposes a software-attack-proof ear recognition system using deep feature learning and blockchain protection to address the problem that the recognition performance of existing systems is generally poor in the face of unconstrained ear databases in realistic scenarios. First, we propose an accommodative DropBlock (AccDrop) to generate drop masks with adaptive shapes. It has an advantage over DropBlock in coping with unconstrained ear databases. Second, we introduce a simple and parameterless attention module that uses 3D weights to refine the ear features output from the convolutional layer. To protect the security of the ear feature template database and the user’s privacy, we use Merkle tree nodes to store the ear feature templates, ensuring the determinism of the root node in the smart contract. We achieve Rank-1 (R1) recognition accuracies of 83.87% and 96.52% on the AWE and EARVN1.0 ear databases, which outperform most advanced ear recognition systems.

Ear images are easy to capture, and ear features are relatively stable and can be used for identification. The ear images are all asymmetric, and the asymmetry of the ear images collected in the unconstrained environment will be more pronounced, increasing the recognition difficulty. Most recognition methods based on hand-crafted features perform poorly in terms of recognition performance in the face of ear databases that vary significantly in terms of illumination, angle, occlusion, and background. This paper proposes a feature fusion human ear recognition method based on channel features and dynamic convolution (CFDCNet). Based on the DenseNet-121 model, the ear features are first extracted adaptively by dynamic convolution (DY_Conv), which makes the ear features of the same class of samples more aggregated and different types of samples more dispersed, enhancing the robustness of the ear feature representation. Then, by introducing an efficient channel attention mechanism (ECA), the weights of important ear features are increased and invalid features are suppressed. Finally, we use the Max pooling operation to reduce the number of parameters and computations, retain the main ear features, and improve the model’s generalization ability. We performed simulations on the AMI and AWE human ear datasets, achieving 99.70% and 72.70% of Rank-1 (R1) recognition accuracy, respectively. The recognition performance of this method is significantly better than that of the DenseNet-121 model and most existing human ear recognition methods.

Sleep, as the basis for regular body functioning, can affect human health. Poor sleep conditions can lead to various physical ailments, such as poor immunity, memory loss, slow cognitive development, and cardiovascular diseases. Along the increasing stress in society comes with a growing surge in conditions associated with sleep disorders. Studies have shown that sleep stages are essential for the body’s memory, immune system, and brain functioning. Therefore, automatic sleep stage classification is of great medical practice importance as a basis for monitoring sleep conditions. Although previous research into the classification of sleep stages has been promising, several challenges remain to be addressed: (1) The EEG signal is a non-smooth signal with harrowing feature extraction and high requirements for model accuracy. (2) Some existing network models suffer from overfitting and gradient descent. (3) Correlation between long time sequences is challenging to capture. This paper proposes NAMRTNet, a deep model architecture based on the original single-channel EEG signal to address these challenges. The model uses a modified ResNet network to extract features from sub-epochs of individual epochs, a lightweight attention mechanism normalization-based attention module (NAM) to suppress insignificant features, and a temporal convolutional network (TCN) network to capture dependencies between features of long time series. The recognition rate of 20-fold cross-validation with the NAMRTNet model for Fpz-cz channel data in the public sleep dataset Sleep-EDF was 86.2%. The experimental results demonstrate the network’s superiority in this paper, surpassing some state-of-the-art techniques in different evaluation metrics. Furthermore, the total time to train the network was 5.1 h, which was much less than the training time of other models.

An extreme learning machine (ELM) is a novel training method for single-hidden layer feedforward neural networks (SLFNs) in which the hidden nodes are randomly assigned and fixed without iterative tuning. ELMs have earned widespread global interest due to their fast learning speed, satisfactory generalization ability and ease of implementation. In this paper, we extend this theory to hypercomplex space and attempt to simultaneously consider multisource information using a hypercomplex representation. To illustrate the performance of the proposed hypercomplex extreme learning machine (HELM), we have applied this scheme to the task of multispectral palmprint recognition. Images from different spectral bands are utilized to construct the hypercomplex space. Extensive experiments conducted on the PolyU and CASIA multispectral databases demonstrate that the HELM scheme can achieve competitive results. The source code together with datasets involved in this paper can be available for free download at https://figshare.com/s/01aef7d48840afab9d6d.

Multispectral palmprint recognition has shown broad prospects for personal identification due to its high accuracy and great stability. In this paper, we develop a novel illumination-invariant multispectral palmprint recognition method. To combine the information from multiple spectral bands, an image-level fusion framework is completed based on a fast and adaptive bidimensional empirical mode decomposition (FABEMD) and a weighted Fisher criterion. The FABEMD technique decomposes the multispectral images into their bidimensional intrinsic mode functions (BIMFs), on which an illumination compensation operation is performed. The weighted Fisher criterion is to construct the fusion coefficients at the decomposition level, making the images be separated correctly in the fusion space. The image fusion framework has shown strong robustness against illumination variation. In addition, a tensor-based extreme learning machine (TELM) mechanism is presented for feature extraction and classification of two-dimensional (2D) images. In general, this method has fast learning speed and satisfying recognition accuracy. Comprehensive experiments conducted on the PolyU multispectral palmprint database illustrate that the proposed method can achieve favorable results. For the testing under ideal illumination, the recognition accuracy is as high as 99.93%, and the result is 99.50% when the lighting condition is unsatisfied.

In this paper, automatic control of the water level in an irrigation canal by automatic regulation of intermediate gates was studied. Previous scholars have proposed a water level difference control strategy that works to keep relative deviations in all pools the same for a particular situation where the operator does not have full control over the canal inflow, with the centralized linear quadratic regulator (LQR) control method used. While in practice, the deviation tolerance of pools may differ in some canals which limits the applicability of the control strategy. In this work, a weight coefficient was added to the deviation and the algorithm was improved to keep the relative deviations to certain proportions. The model predictive control (MPC) method was then used with this improved control strategy and was compared to the LQR control method using the same control strategy. The results showed that the improved strategy can keep the water level deviations in all pools to certain proportions, as is our objective. Also, under this difference control strategy, the MPC method greatly improved the control performance compared to the LQR control method.

Peak discharge is an essential element of hydrological forecasting. Due to rapid outbreaks of flash floods in hilly areas and the lack of measured data, the fast and accurate estimation of peak discharge is crucial for flash flood hazard management. Three machine learning algorithms were applied to estimate peak discharge; this estimation was compared with the results of hydrological--hydraulic models, and the results were verified with measured watershed data. In this paper, 10 hydrological and geomorphological parameters were selected to predict the flood peak discharge in 103 watersheds in Taiyi Mountain North District. The results show that the particle swarm optimization backpropagation (PSO-BP) neural network model outperforms the BP neural network and random forest regression in prediction performance. PSO-BP has a lower mean absolute error (2.51%), root mean square error (3.74%), and mean absolute percentage error (2.74%) than the other models, which indicates that PSO-BP has high prediction accuracy. Importance analysis revealed that rainfall, early impact rainfall, catchment area, and rain intensity are the key input parameters of PSO-BP. The proposed method was confirmed to be a fast and relatively accurate algorithm for estimating the peak discharge of flash floods in ungauged basins. [ABSTRACT FROM AUTHOR]

To achieve accurate detection of tuberculosis (TB) areas in chest radiographs, we design a chest X-ray TB area detection algorithm. The algorithm consists of two stages: the chest Xray TB classification network (CXTCNet) and the chest X-ray TB area detection network (CXTDNet). CXTCNet is used to judge the presence or absence of TB areas in chest X-ray images, thereby excluding the influence of other lung diseases on the detection of TB areas. It can reduce false positives in the detection network and improve the accuracy of detection results. In CXTCNet, we propose a channel attention mechanism (CAM) module and combine it with DenseNet. This module enables the network to learn more spatial and channel features information about chest X-ray images, thereby improving network performance. CXTDNet is a design based on a sparse object detection algorithm (Sparse R-CNN). A group of fixed learnable proposal boxes and learnable proposal features are using for classification and location. The predictions of the algorithm are output directly without non-maximal suppression post-processing. Furthermore, we use CLAHE to reduce image noise and improve image quality for data preprocessing. Experiments on dataset TBX11K show that the accuracy of the proposed CXTCNet is up to 99.10%, which is better than most current TB classification algorithms. Finally, our proposed chest X-ray TB detection algorithm could achieve AP of 45.35% and AP50 of 74.20%. We also establish a chest X-ray TB dataset with 304 sheets. And experiments on this dataset showed that the accuracy of the diagnosis was comparable to that of radiologists. We hope that our proposed algorithm and established dataset will advance the field of TB detection. [ABSTRACT FROM AUTHOR]

Lip-reading techniques have shown bright prospects for speech recognition under noisy environments and for hearing-impaired listeners. We aim to solve two important issues regarding lip reading: (1) how to extract discriminative lip motion features and (2) how to establish a classifier that can provide promising recognition accuracy for lip reading. For the first issue, a projection local spatiotemporal descriptor, which considers the lip appearance and motion information at the same time, is utilized to provide an efficient representation of a video sequence. For the second issue, a kernel extreme learning machine (KELM) based on the single-hiddenlayer feedforward neural network is presented to distinguish all kinds of utterances. In general, this method has fast learning speed and great robustness to nonlinear data. Furthermore, quantum-behaved particle swarm optimization with binary encoding is introduced to select the appropriate feature subset and parameters for KELM training. Experiments conducted on the AVLetters and OuluVS databases show that the proposed lipreading method achieves a superior recognition accuracy compared with two previous methods. [ABSTRACT FROM AUTHOR]