Multi-branch Structure Research Articles

Multi-Source Information-Based Bearing Fault Diagnosis Using Multi-Branch Selective Fusion Deep Residual Network.

Rolling bearing is the core component of industrial machines, but it is difficult for common single signal source-based fault diagnosis methods to ensure reliable results since sensor signals are vulnerable to the pollution of background noises and the attenuation of transmitted information. Recently, multi-source information-based fault diagnosis methods have become popular, but the information redundancy between multiple signals is a tough problem that will negatively impact the representational capacity of deep learning algorithms and the precision of fault diagnosis methods. Besides that, the characteristics of various signals are actually different, but this problem was usually omitted by researchers, and it has potential to further improve the diagnosing performance by adaptively adjusting the feature extraction process for every input signal source. Aimed at solving the above problems, a novel model for bearing fault diagnosis called multi-branch selective fusion deep residual network is proposed in this paper. The model adopts a multi-branch structure design to enable every input signal source to have a unique feature processing channel, avoiding the information of multiple signal sources blindly coupled by convolution kernels. And in each branch, different convolution kernel sizes are assigned according to the characteristics of every input signal, fully digging the precious fault components on respective information sources. Lastly, the dropout technique is used to randomly throw out some activated neurons, alleviating the redundancy and enhancing the quality of the multiscale features extracted from different signals. The proposed method was experimentally compared with other intelligent methods on two authoritative public bearing datasets, and the experimental results prove the feasibility and superiority of the proposed model.

End-to-end color fringe depth estimation based on a three-branch U-net network

In fringe projection profilometry (FPP), end-to-end depth estimation from fringe patterns for FPP attracts more and more attention from fringe patterns. However, color images provide additional information from the RGB channel for FPP, which has been paid little attention in depth estimation. To this end, in this paper we present for the first time, to the best of our knowledge, an end-to-end network for depth estimation using color composite fringes with better performance. In order to take advantage of the color fringe pattern, a multi-branch structure is designed in this paper, which learns the multi-channel details of the object under test by using three encoders for each RGB channel and introduces an attention module to better capture the complex features and modalities information in the input data. Experiments from simulated and real datasets show that the proposed method with color fringe pattern is effective for depth estimation, and it outperforms other deep learning methods such as UNet, R2Unet, PCTNet, and DNCNN.

Pyrolysis kinetics and flame retardant enhancement of bio-based polyamide 56/6

The development of polyamide materials with fire safety is of great importance at this stage. A novel nitrogen-phosphorus bisystem flame retardant (MC) with a multi-branched structure was synthesized and applied to a new bio-based polyamide 56/6 (PA56/6). Notably, at 8 wt% MC content, flame-retardant PA56/6@MC8% (FRPA56/6@MC8%) achieved an Limiting Oxygen Index (LOI) of 26.6% and a V-0 rating in UL-94 tests. Cone calorimetry results indicated that FRPA56/6@MC8% exhibited a 22.9% reduction in total heat release (THR) and a 41.0% decrease in peak heat release rate (PHRR), underscoring the flame retardancy promotion by MC in PA56/6. The study further explored the pyrolysis kinetics and mechanisms of polyamide materials, offering insights crucial for flame-retardant modifications. Overall, the findings present an innovative strategy for enhancing the flame retardant properties of PA56/6, potentially applicable in automotive components and other pertinent fields.

Aileron fault detection with dynamic resampling based on fuzzy entropy and multi-branch neural network

The complex movements of unmanned aerial vehicles (UAVs) and feature loss at low resampling rates will result in the misidentification of aileron damage. In this paper, a fault detection method based on fuzzy entropy and a multi-branch neural network (MBNN) is proposed to shorten the fault detection time by reducing the resampling rate. The optimal size of the sliding window sampler is determined by analysing the UAV’s movement using fuzzy entropy, ensuring that the captured flight data segments reflect the UAV’s flight status with a smaller size. The multi-branch structure and feature-sharing mechanism in the network are used to adapt to feature loss at low sampling rates. The resampling rate is dynamically adjusted in accordance with the change in the UAV’s fault status, which is calculated from MBNN’s output using the Kullback Leibler (KL) divergence. Experimental results show that the proposed method can improve the real-time performance of UAV fault detection and maintain high accuracy.

3D morphological evolution of the Fe-Rich phase and mechanical properties of the recyclable A356 alloy

Efficient utilisation of recycled aluminium promotes low-carbon and green manufacturing in the aluminium industry. However, the low tolerance of Al for Fe content limits its application. The morphological evolution of the Fe-rich phase and the mechanical properties of recycled A356 alloys with varying Fe contents were investigated using optical microscopy (OM), scanning electron microscopy (SEM), differential scanning calorimetry (DSC), synchrotron radiation X-ray computed tomography (SRXCT), and Thermo-Calc thermodynamic calculations. The results show that with increasing Fe content, the morphology of the Fe-rich phase transformed from fine-granular and short rod-like structures into coarse needle-like and Chinese-script. In three dimensions (3D), the larger Fe-rich phases displayed a multibranched structure composed of rod-like and plate-like phases rather than a single morphology. Notably, at an Fe content of 0.4 %, the Fe-rich phases presented Chinese-script and short rod-like shapes in sections but did not significantly change the 3D morphology or type of the Fe-rich phase. However, the Fe-rich phases in the 0.4 %Fe alloy exhibited higher thickness and sphericity, which improved the ductility of the alloy. Thermodynamic calculations indicate that the initial formation temperature of the Fe-rich phase in the 0.4 % Fe alloy overlaps with that of the Al-Si eutectic reaction, which is referred to as the critical Fe content. At the critical Fe content, the ternary eutectic reaction of Al-Si-(AlFeSi) suppressed the rapid growth of the Fe-rich phase and mitigated their negative impact on ductility.

Abnormal behavior detection in industrial control systems based on CNN

With the widespread application of Internet of Things technology in industrial control systems, abnormal behavior detection has become a key task to ensure the safety and stable operation of the system. We propose a multi-branch convolutional fusion neural network method to improve the accuracy and efficiency of abnormal behavior detection in Internet of Things industrial control systems. This method achieves efficient detection of abnormal behavior in video data by combining ResNet152 for spatial feature extraction and GRU for temporal feature extraction. Unlike traditional methods, this method adopts a multi-branch structure, which can simultaneously capture multi-scale feature information, significantly enhancing the richness of feature expression and the accuracy of detection. Experimental results show that on the UCF-Crime dataset, the accuracy of this method reaches 85.76%, which is significantly better than that of traditional methods. In addition, on the larger UCF-101 dataset, the accuracy of this method reaches 92.21%, further verifying its excellent generalization performance. Compared with the C3D network, this method improves the accuracy by nearly 6% while maintaining a high processing speed. These results show that the proposed method has great potential in practical applications.

Research on Rail Surface Defect Detection Based on Improved CenterNet

Rail surface defect detection is vital for railway safety. Traditional methods falter with varying defect sizes and complex backgrounds, while two-stage deep learning models, though accurate, lack real-time capabilities. To overcome these challenges, we propose an enhanced one-stage detection model based on CenterNet. We replace ResNet with ResNeXt and implement a multi-branch structure for better low-level feature extraction. Additionally, we integrate SKNet attention mechanism with the C2f structure from YOLOv8, improving the model’s focus on critical image regions and enhancing the detection of minor defects. We also introduce an elliptical Gaussian kernel for size regression loss, better representing the aspect ratio of rail defects. This approach enhances detection accuracy and speeds up training. Our model achieves a mean accuracy (mAP) of 0.952 on the rail defects dataset, outperforming other models with a 6.6% improvement over the original and a 35.5% increase in training speed. These results demonstrate the efficiency and reliability of our method for rail defect detection.

Fast Semantic Segmentation of Ultra-High-Resolution Remote Sensing Images via Score Map and Fast Transformer-Based Fusion

For ultra-high-resolution (UHR) image semantic segmentation, striking a balance between computational efficiency and storage space is a crucial research direction. This paper proposes a Feature Fusion Network (EFFNet) to improve UHR image semantic segmentation performance. EFFNet designs a score map that can be embedded into the network for training purposes, enabling the selection of the most valuable features to reduce storage consumption, accelerate speed, and enhance accuracy. In the fusion stage, we improve upon previous redundant multiple feature fusion methods by utilizing a transformer structure for one-time fusion. Additionally, our combination of the transformer structure and multibranch structure allows it to be employed for feature fusion, significantly improving accuracy while ensuring calculations remain within an acceptable range. We evaluated EFFNet on the ISPRS two-dimensional semantic labeling Vaihingen and Potsdam datasets, demonstrating that its architecture offers an exceptionally effective solution with outstanding semantic segmentation precision and optimized inference speed. EFFNet substantially enhances critical performance metrics such as Intersection over Union (IoU), overall accuracy, and F1-score, highlighting its superiority as an architectural innovation in ultra-high-resolution remote sensing image semantic segmentation.

A Multi-Branch Feature Extraction Residual Network for Lightweight Image Super-Resolution

Single-image super-resolution (SISR) seeks to elucidate the mapping relationships between low-resolution and high-resolution images. However, high-performance network models often entail a significant number of parameters and computations, presenting limitations in practical applications. Therefore, prioritizing a light weight and efficiency becomes crucial when applying image super-resolution (SR) to real-world scenarios. We propose a straightforward and efficient method, the Multi-Branch Feature Extraction Residual Network (MFERN), to tackle lightweight image SR through the fusion of multi-information self-calibration and multi-attention information. Specifically, we have devised a Multi-Branch Residual Feature Fusion Module (MRFFM) that leverages a multi-branch residual structure to succinctly and effectively fuse multiple pieces of information. Within the MRFFM, we have designed the Multi-Scale Attention Feature Fusion Block (MAFFB) to adeptly extract features via convolution and self-calibration attention operations. Furthermore, we introduce a Dual Feature Calibration Block (DFCB) to dynamically fuse feature information using dynamic weight values derived from the upper and lower branches. Additionally, to overcome the limitation of convolution in solely extracting local information, we incorporate a Transformer module to effectively integrate global information. The experimental results demonstrate that our MFERN exhibits outstanding performance in terms of model parameters and overall performance.

Research on Tool Wear Monitoring Based on Enhanced Convolutional Neural Networks

Abstract Identifying the wear status of cutting tools during the machining process is essential because failure to promptly replace severely worn tools can significantly impact the quality of workpiece machining. Presently, machine learning methods are predominantly utilized for monitoring cutting tool wear status. However, these methods rely on manual feature extraction and exhibit low accuracy. This study introduces a novel RRP-Net model, built upon the RepVgg and ResNet frameworks, integrating a parameter-free self-attention mechanism called SimAM to expedite the model’s solving speed without increasing parameters. Within the foundational module of the model, a structural reparameterization approach is employed to transform the multi-branch structure during training into a single-branch structure during validation. This method not only enhances model accuracy but also accelerates the model validation process. The publicly available cutting data from PHM2010 is employed for model training and validation. The findings demonstrate that RRP-Net surpasses classical convolutional neural network models in identifying cutting tool wear status within the PHM2010 dataset, achieving an average accuracy of 98.65% and enhancing recognition accuracy on relevant datasets by 2.41%. To verify the model’s practical applicability, specificity and recall during the Break stage are computed at 99.73% and 98.18%, respectively, affirming the model’s exceptional robustness and stability. The heightened accuracy and efficiency of RRP-Net further broaden its applicability within the industrial domain.

SMR–YOLO: Multi-Scale Detection of Concealed Suspicious Objects in Terahertz Images

The detection of concealed suspicious objects in public places is a critical issue and a popular research topic. Terahertz (THz) imaging technology, as an emerging detection method, can penetrate materials without emitting ionizing radiation, providing a new approach to detecting concealed suspicious objects. This study focuses on the detection of concealed suspicious objects wrapped in different materials such as polyethylene and kraft paper, including items like scissors, pistols, and blades, using THz imaging technology. To address issues such as the lack of texture details in THz images and the contour similarity of different objects, which can lead to missed detections and false alarms, we propose a THz concealed suspicious object detection model based on SMR–YOLO (SPD_Mobile + RFB + YOLO). This model, based on the MobileNext network, introduces the spatial-to-depth convolution (SPD-Conv) module to replace the backbone network, reducing computational and parameter load. The inclusion of the receptive field block (RFB) module, which uses a multi-branch structure of dilated convolutions, enhances the network’s depth features. Using the EIOU loss function to assess the accuracy of predicted box localization further optimizes convergence speed and localization accuracy. Experimental results show that the improved model achieved mAP@0.5 and mAP@0.5:0.95 scores of 98.9% and 89.4%, respectively, representing improvements of 0.2% and 1.8% over the baseline model. Additionally, the detection speed reached 108.7 FPS, an improvement of 23.2 FPS over the baseline model. The model effectively identifies concealed suspicious objects within packages, offering a novel approach for detection in public places.

Coadsorption mechanisms of copper and sulfamethoxazole by functionalized cellulose: A kinetic and DFT study

In order to address the issue of compound pollution from heavy metals and antibiotics in aquatic environments, multibranched functionalized cellulose materials (DACS) with abundant adsorption affinity groups (such as amino, carboxyl and imine groups) were prepared by oxidation, etherification and grafting. The maximum adsorption capacities of DACS for copper (Cu) and sulfamethoxazole (SMZ) were 42.27 and 117.92 mg/g, respectively. In addition, its excellent pH buffering capacity, resistance to coexisting interfering substances and great regeneration performance of up to 91 % after 5 cycles. In the coadsorption experiment, the multibranched structure strengthened the bridge effect during the coadsorption process, and weakened the inhibition of SMZ adsorption caused by competition. Kinetic models and error functions analysis showed that PNO model and F&S model were the most consistent with the adsorption process, and the limiting step of adsorption was the external diffusion of Cu and SMZ around the surface of DACS. Moreover, the density functional theory (DFT) was used to quantitatively calculate the adsorption binding energy of multiple sites on DACS, and clarified the adsorption tendency of DACS as electron donors for Cu and as acceptor for SMZ. In addition, the binding energy of sequential adsorption was greater than that of single and binary adsorption, and the binding stability of SMZ was significantly increased in sequential adsorption for the increasing of binding energy. The electron exchange process in the adsorption showed that SMZ played a bridging role in the binary adsorption, while Cu was the main reaction part of the complex adsorption.

A Robust Multi-Camera Vehicle Tracking Algorithm in Highway Scenarios Using Deep Learning

In intelligent traffic monitoring systems, the significant distance between cameras and their non-overlapping fields of view leads to several issues. These include incomplete tracking results from individual cameras, difficulty in matching targets across multiple cameras, and the complexity of inferring the global trajectory of a target. In response to the challenges above, a deep learning-based vehicle tracking algorithm called FairMOT-MCVT is proposed. This algorithm con-siders the vehicles’ characteristics as rigid targets from a roadside perspective. Firstly, a Block-Efficient module is designed to enhance the network’s ability to capture and characterize image features across different layers by integrating a multi-branch structure and depth-separable convolutions. Secondly, the Multi-scale Dilated Attention (MSDA) module is introduced to improve the feature extraction capability and computational efficiency by combining multi-scale feature fusion and attention mechanisms. Finally, a joint loss function is crafted to better distinguish between vehicles with similar appearances by combining the trajectory smoothing loss and velocity consistency loss, thereby considering both position and velocity continuity during the optimization process. The proposed method was evaluated on the public UA-DETRAC dataset, which comprises 1210 video sequences and over 140,000 frames captured under various weather and lighting conditions. The experimental results demonstrate that the FairMOT-MCVT algorithm significantly enhances multi-target tracking accuracy (MOTA) to 79.0, IDF1 to 84.5, and FPS to 29.03, surpassing the performance of previous algorithms. Additionally, this algorithm expands the detection range and reduces the deployment cost of roadside equipment, effectively meeting the practical application requirements.

An improved attentive residue multi-dilated network for thermal noise removal in magnetic resonance images

Magnetic resonance imaging (MRI) technology is crucial in the medical field, but the thermal noise in the reconstructed MR images may interfere with the clinical diagnosis. Removing the thermal noise in MR images mainly contains two challenges. First, thermal noise in an MR image obeys Rician distribution, where the statistical features are not consistent in different regions of the image. In this case, conventional denoising methods like spatial convolutional filtering will not be appropriate to deal with it. Second, details and edge information in the image may get damaged while smoothing the noise. This paper proposes a novel deep-learning model to denoise MR images. First, the model learns a binary mask to separate the background and signal regions of the noised image, making the noise left in the signal region obey a unified statistical distribution. Second, the model is designed as an attentive residual multi-dilated network (ARM-Net), composed of a multi-branch structure, and supplemented with a frequency-domain-optimizable discrete cosine transform module. In this way, the deep-learning model will be more effective in removing the noise while maintaining the details of the original image. Furthermore, we have also made improvements on the original ARM-Net baseline to establish a new model called ARM-Net v2, which is more efficient and effective. Experimental results illustrate that over the BraTS 2018 dataset, our method achieves the PSNR of 39.7087 and 32.6005 at noise levels of 5% and 20%, which realizes the state-of-the-art performance among existing MR image denoising methods.

A Novel Real-Time Detection and Classification Method for ECG Signal Images Based on Deep Learning.

In this paper, a novel deep learning method Mamba-RAYOLO is presented, which can improve detection and classification in the processing and analysis of ECG images in real time by integrating three advanced modules. The feature extraction module in our work with a multi-branch structure during training can capture a wide range of features to ensure efficient inference and rich feature extraction. The attention mechanism module utilized in our proposed network can dynamically focus on the most relevant spatial and channel-wise features to improve detection accuracy and computational efficiency. Then, the extracted features can be refined for efficient spatial feature processing and robust feature fusion. Several sets of experiments have been carried out to test the validity of the proposed Mamba-RAYOLO and these indicate that our method has made significant improvements in the detection and classification of ECG images. The research offers a promising framework for more accurate and efficient medical ECG diagnostics.

Synthesis and growth mechanism of highly crystalized multi-branched HfB2 microrods with self-toughening effect

Development of strategies to toughen HfB2 ceramic, improve their mechanical properties is the key to promote their applications in extreme thermal environments. This study explores a kind of self-toughening method of HfB2 ceramic by incorporation the highly crystalized multi-branched HfB2 microrods. The multi-branched HfB2 microrods were synthesized by sol–gel process and carbothermal reduction, exhibiting length of 6–13 μm and diameter of 0.60–1.50 μm, respectively. Detailed analysis of the formation mechanism of HfB2 microrods reveals that the HfB2 multi-branched structure is a collaborative consequence of orientation attachment mechanism and reactive templates growth, in which process monoclinic HfO2 crystals transform into crystalized HfO2 microrods at ∼1000 °C, and worked as template to induce one-dimensional HfB2 microrods. Furthermore, spark plasma sintering experiments show that the optimal mechanical properties were attained by 6 wt% addition of HfB2 microrods, comparing with HfB2 ceramics free of self-toughening agent, the hardness and fracture toughness of self-toughened HfB2 ceramic improved by 8.82 % and 81.91 % respectively, demonstrating the effectiveness of self-toughening of the microrods without sacrificing the hardness of HfB2 ceramic. Thus, it can be seen that the work has important reference value for the synthesis of HfB2 powder with different morphologies and the toughening of HfB2 ceramics.

RNE-DSNet: A Re-parameterization Neighborhood Enhancement-based Dual-Stream Network for CT image recognition

Deep learning models based on Transformer and CNN are current research hotspots. However, there are more time complexity and space complexity in extracting global–local features, the feature extraction ability is insufficient for small lesions in CT images. A Re-parameterization Neighborhood Enhancement-based Dual-Stream Network (RNE-DSNet) is proposed for CT image recognition in this paper. This method improves the ability to recognize small lesions while reducing the Params number and FLOPs, which gives a trade-off between performance and cost. The main works are as following: Firstly, a Re-parameterization Neighborhood Enhancement-based Dual-Stream Network is designed to improve the global–local feature extraction ability. Secondly, the Re-parameterization Group Residual Block is designed in the CNN branch, which improves the local feature extraction ability through multi-scale and multi-branch structures, and the number of the Params and FLOPs are reduced effectively. The Neighborhood Enhancement Transformer Block is designed in the Transformer branch, in which the global attention is calculated by parallel Hadamard product and Dot product, and the global attention maps are enhanced by deep convolution operation to embed local features in global features. Thirdly, the Adaptive-random Mask Cropping method is designed to improve the features extraction ability for small lesions in CT images through heat map and random Mask techniques. RNE-DSNet is compared with other SOTA models. The Accuracy (ACC), Precision (PRE), Recall (REC), F1 score (F1), and Area Under Curve (AUC) values of RNE-DSNet are reached 99.22%, 99.23%, 99.08%, 99.23% and 99.38%, respectively, which verified the effectiveness of the RNE-DSNet model. The number Params and FLOPs are reduced by 62.78% and 84.20%, respectively. The effectiveness of RNE-DSNet is verified by heat map visualization. The performance of RNE-DSNet is generally better than the others methods. There is positive significance for computer aided diagnosis.

Res‐MulFra: Multilevel and Multiscale Framework for Brain Tumor Segmentation

ABSTRACTIn clinical diagnosis and surgical planning, extracting brain tumors from magnetic resonance images (MRI) is very important. Nevertheless, considering the high variability and imbalance of the brain tumor datasets, the way of designing a deep neural network for accurately segmenting the brain tumor still challenges the researchers. Moreover, as the number of convolutional layers increases, the deep feature maps cannot provide fine‐grained spatial information, and this feature information is useful for segmenting brain tumors from the MRI. Aiming to solve this problem, a brain tumor segmenting method of residual multilevel and multiscale framework (Res‐MulFra) is proposed in this article. In the proposed framework, the multilevel is realized by stacking the proposed RMFM‐based segmentation network (RMFMSegNet), which is mainly used to leverage the prior knowledge to gain a better brain tumor segmentation performance. The multiscale is implemented by the proposed RMFMSegNet, which includes both the parallel multibranch structure and the serial multibranch structure, and is mainly designed for obtaining the multiscale feature information. Moreover, from various receptive fields, a residual multiscale feature fusion module (RMFM) is also proposed to effectively combine the contextual feature information. Furthermore, in order to gain a better brain tumor segmentation performance, the channel attention module is also adopted. Through assessing the devised framework on the BraTS dataset and comparing it with other advanced methods, the effectiveness of the Res‐MulFra is verified by the extensive experimental results. For the BraTS2015 testing dataset, the Dice value of the proposed method is 0.85 for the complete area, 0.72 for the core area, and 0.62 for the enhanced area.

Highly Accurate and Lightweight Detection Model of Apple Leaf Diseases Based on YOLO

To mitigate problems concerning small-sized spots on apple leaves and the difficulties associated with the accurate detection of spot targets exacerbated by the complex backgrounds of orchards, this research used alternaria leaf spots, rust, brown spots, gray spots, and frog eye leaf spots on apple leaves as the research object and proposed the use of a high-accuracy detection model YOLOv5-Res (YOLOv5-Resblock) and lightweight detection model YOLOv5-Res4 (YOLOv5-Resblock-C4). Firstly, a multiscale feature extraction module, ResBlock (residual block), was designed by combining the Inception multi-branch structure and ResNet residual idea. Secondly, a lightweight feature fusion module C4 (CSP Bottleneck with four convolutions) was designed to reduce the number of model parameters while improving the detection ability of small targets. Finally, a parameter-streamlining strategy based on an optimized model architecture was proposed. The experimental results show that the performance of the YOLOv5-Res model and YOLOv5-Res4 model is significantly improved, with the mAP0.5 values increasing by 2.8% and 2.2% compared to the YOLOv5s model and YOLOv5n model, respectively. The sizes of the YOLOv5-Res model and YOLOv5-Res4 model are only 10.8 MB and 2.4 MB, and the model parameter counts are reduced by 22% and 38.3% compared to the YOLOv5s model and YOLOv5n model.

Comprehensive analysis of band gap modulation of hexagonal fan blade and optimized ligament structure in the low-frequency range

This paper proposed a meta material model with low-frequency wide band gap and band gap tunability, optimized its structure as a multibranch chain structure, and analyzed the band gap relationship of above two models based on Brag's theorem and finite element method. Among them, the fan base structure has a band gap of 65.92 % within 20,000 Hz, and the band gap is optimized to low-frequency after the structure is optimized, and the band gap reaches 75.94 % within 10,000 Hz. The effects of the ligament width and the thickness of the center parcel layer on the band gap distribution of the structure are also discussed, and the wave transmission characteristics in the structure are explored by group and phase velocities to verify the acoustic performance of the structure. The results show that the smaller the ligament thickness is, the lower the first band gap opening frequency of the structure is; when the thickness of the central wrapping layer is increased, the structure develops the band gap at the center frequency of 3000 Hz to a lower frequency and generates an ultra-wide band gap at the center frequency of 6000 Hz. These findings can provide a new idea for low-frequency vibration isolation and noise reduction.