Fusion Operation Research Articles

Reference-based image super-resolution of hyperspectral and red-green-blue image for determination of wheat kernel quality using deep learning networks

In the process of cultivation and harvest, wheat kernel quality is highly susceptible to various factors, such as disease, mildew, atrophy and impurities, and detection of kernel quality is essential to avoid hazard proliferation, facilitate product grading, and ensure food safety. Possessing abundant image and spectral characteristics, hyperspectral imaging (HSI) has gained impressive achievements in kernel quality analysis, but its low spatial resolution limits its detection accuracy. In this study, reference-based image super-resolution (RefSR) of HSI and Red-Green-Blue image was adopted to improve resolution to determine wheat kernel quality using deep learning networks. Firstly, RefSR was conducted by the improved transformer network with dual-branch feature extraction and weighted fusion operation and achieved excellent RefSR with significant resolution improvement, peak signal to noise ratio of 35.521 and structural similarity index of 0.97, outweighing the existing state-of-the-art networks. Then, the reflectance images (RIs) of effective wavelengths (EWs) from generated HSI images were combined with the residual network with a spatial, channel attention and multi-scale residual to determine wheat kernel quality. Precise analysis was achieved with the accuracy in calibration, validation and prediction sets of 100.00%, 95.26% and 92.78%. RefSR provides a novel and efficient approach for obtaining HSI images of high spatial resolution and facilitates the application of HSI in analysis of crop kernels. RIs of several sporadic EWs can be easily acquired and processed, achieving field and rapid kernel detection. Therefore, the proposed method furnishes the efficient, accurate and applicable determination of wheat kernel quality.

Graph-induced rank-aggregation using information fusion operators

Graph-induced rank-aggregation using information fusion operators

Widely wavelength-tunable high-repetition-rate femtosecond pulse source with highest average power up to 28 W

We have proposed and demonstrated a high-repetition-rate ultrashort pulse fiber amplification system based on a wavelength-tunable oscillator. This fiber amplification system produces an average power exceeding 20 W in bursts of 200 pulses with a 578 MHz intra-burst pulse repetition rate and a 1 MHz burst repetition rate. The center wavelength of the amplified pulses can be tuned from 1030 to 1080 nm. By utilizing pre-chirp management nonlinear amplification technique, the achieved shortest pulse duration is 27 fs with an average power of 25 W at 1032 nm. For all the amplified pulses with different wavelengths, the pulse duration after optimal compression is below 60 fs. To the best of our knowledge, this is the first time that a widely wavelength-tunable high-power laser with a repetition rate exceeding 100 MHz and a pulse duration of several tens of femtoseconds has been realized. Additionally, using only common double-cladding Yb-doped fiber as the gain fiber, without any large-mode-area Yb-doped photonic crystal fiber or rod-type Yb-doped fiber, makes the system compact and reliable due to the simple fusion operation.

Multimodal Deep Representation Learning for Quantum Cross-Platform Verification.

Cross-platform verification, a critical undertaking in the realm of early-stage quantum computing, endeavors to characterize the similarity of two imperfect quantum devices executing identical algorithms, utilizing minimal measurements. While the random measurement approach has been instrumental in this context, the quasiexponential computational demand with increasing qubit count hurdles its feasibility in large-qubit scenarios. To bridge this knowledge gap, here we introduce an innovative multimodal learning approach, recognizing that the formalism of data in this task embodies two distinct modalities: measurement outcomes and classical description of compiled circuits on explored quantum devices, both containing unique information about the quantum devices. Building upon this insight, we devise a multimodal neural network to independently extract knowledge from these modalities, followed by a fusion operation to create a comprehensive data representation. The learned representation can effectively characterize the similarity between the explored quantum devices when executing new quantum algorithms not present in the training data. We evaluate our proposal on platforms featuring diverse noise models, encompassing system sizes up to 50 qubits. The achieved results demonstrate an improvement of 3 orders of magnitude in prediction accuracy compared to the random measurements and offer compelling evidence of the complementary roles played by each modality in cross-platform verification. These findings pave the way for harnessing the power of multimodal learning to overcome challenges in wider quantum system learning tasks.

An hetero-modal deep learning framework for medical image synthesis applied to contrast and non-contrast MRI

Some pathologies such as cancer and dementia require multiple imaging modalities to fully diagnose and assess the extent of the disease. Magnetic resonance imaging offers this kind of polyvalence, but examinations take time and can require contrast agent injection. The flexible synthesis of these imaging sequences based on the available ones for a given patient could help reduce scan times or circumvent the need for contrast agent injection. In this work, we propose a deep learning architecture that can perform the synthesis of all missing imaging sequences from any subset of available images. The network is trained adversarially, with the generator consisting of parallel 3D U-Net encoders and decoders that optimally combines their multi-resolution representations with a fusion operation learned by an attention network trained conjointly with the generator network. We compare our synthesis performance with 3D networks using other types of fusion and a comparable number of trainable parameters, such as the mean/variance fusion. In all synthesis scenarios except one, the synthesis performance of the network using attention-guided fusion was better than the other fusion schemes. We also inspect the encoded representations and the attention network outputs to gain insights into the synthesis process, and uncover desirable behaviors such as prioritization of specific modalities, flexible construction of the representation when important modalities are missing, and modalities being selected in regions where they carry sequence-specific information. This work suggests that a better construction of the latent representation space in hetero-modal networks can be achieved by using an attention network.

Fusion of Multiple Attention Mechanisms and Background Feature Adaptive Update Strategies in Siamese Networks for Single-Object Tracking

Single-object tracking algorithms based on Siamese full convolutional networks have attracted much attention from researchers owing to their improvement in precision and speed. Since this tracking model only learns a similarity model offline, it is not able to obtain more useful feature discrimination information to adapt to the various variations of targets in complex scenes. To improve the performance of this tracking model, we propose a Siamese network tracking algorithm that incorporates multiple attention mechanisms and an adaptive updating strategy for background features. First, a backbone feature extraction network is proposed that utilizes a small convolutional kernel to fuse jump-layer connectivity features, thereby improving the feature representation capability of the network. Second, an adaptive update strategy for background features is proposed to improve the model’s ability to discriminate between the object and background features. Third, the fusion of multiple attention mechanisms is proposed so that the model learns to focus on the channel, spatial, and coordinate features. Fourth, the response fusion operation is proposed after the inter-correlation operation to enrich the output response of the model. Finally, our algorithm is trained using the GOT-10K dataset and evaluated by testing on the object tracking benchmark datasets OTB100 and VOT2018. The test results show that compared with other algorithms, our algorithm can effectively cope with the problem of degradation of the tracking performance in complex environments, and it can further improve the tracking precision and precision under the premise of ensuring the tracking speed.

Design concept of intelligent integrated control system for neutral beam injection

Abstract Due to the specificity of neutral beam injection (NBI) system, the control of its actual physical system is achieved by the integrated control system (ICS). The current NBI ICS adopts a distributed design structure to balance the system load, which is also the mainstream system design and architecture model of ICS around the world. However, from a practical point of view, the distributed system architecture is not perfect. In the long run, upgrading the system intelligence and automation capabilities is an inevitable choice for the future, so it is necessary to explore ways to transform and upgrade the ICS. At present, Internet of Things (IoT), as an important part of the new generation information technology, can realize the control of everything through data exchange. This is because on the one hand, IoT has wide compatibility and powerful scenario-based capabilities, it not only has the advantages and characteristics of distributed design, but also can pull the NBI subsystems into the same level scenario and lay the foundation for the further construction of digital NBI; on the other hand, the intervention of artificial intelligence makes IoT have some new typical characteristics such as intelligent sensing, ubiquitous connectivity, precise control, digital modeling, real-time analysis and iterative optimization, which is enough to pull the current NBI ICS into a new intelligent control era. Finally, it is worth mentioning that due to its inherent design structure and functional characteristics, ICS tends to be broadly generic, so it is not exclusively used for NBI operation in nuclear fusion, and it can provide some insight into other application areas.

Patient acceptance of reoperation risk for lumbar decompression versus fusion

BACKGROUND CONTEXTLumbar decompression and lumbar fusion are effective methods of treating spinal compressive pathologies refractory to conservative management. These surgeries are typically used to treat different spinal problems, but there is a growing body of literature investigating the outcomes of either approach for patients with lumbar degenerative spondylolisthesis and stenosis. Different operations are associated with different risks and different potential needs for reoperation. Patient acceptance of reoperation rates after spinal surgery is currently not well understood. PURPOSEThe purpose of this study is to identify patient tolerance for reoperation rates following lumbar decompression and lumbar fusion surgery. DESIGNA qualitative and quantitative survey intended to capture information on patient preferences was administered. PATIENT SAMPLEWritten informed consent was obtained from patients presenting to 2 spinal clinics. OUTCOME MEASURESPatients were asked their threshold tolerance for reoperation rates in the context of choosing a smaller (decompression) versus larger (fusion) spinal surgery. METHODSA survey was administered to patients at 2 spinal clinics—1 surgical and 1 nonsurgical. A consecutive series of new patients over multiple clinic days who agreed to participate in the study and filled out the survey are reported on here. Patients were asked to assess, contemplating a problem that could either be treated with lumbar decompression or lumbar fusion, the level at which 1) the likelihood that needing a repeat surgery within 3 to 5 years would change their mind about choosing the decompression operation and cause them to choose the fusion operation and then 2) the likelihood of needing a repeat surgery within 3 to 5 years that would be acceptable to them after the fusion operation. The distribution of patient responses was assessed with histograms and descriptive statistics. RESULTSNinety patients were surveyed, and of these, 73 patients (81.1%) returned fully completed questionnaires. The median reoperation acceptance rates after a decompression was <60%, while the median acceptable revision rate when contemplating the fusion surgery was 10%. CONCLUSIONSPatient acceptance for the potential need for revision surgery is higher when considering a decompression compared to a fusion operation. Reoperation risk rates along with the magnitude of the surgical intervention are important considerations in determining patients’ surgical preferences. Understanding patient preferences and risk tolerances can aid clinicians in shared decision-making, potentially improving patient satisfaction and outcomes in the several lumbar pathologies which can be ameliorated with either decompression or fusion.

Hotel recommendation mechanism based on online reviews considering multi-attribute cooperative and interactive characteristics

Online reviews of hotels provide important information to consumers. The process of extracting useful information from diverse online reviews is crucial for making the best final decisions. To explore the hidden intrinsic information behind online reviews, this paper optimizes information extraction by integrating multiple sources, and gives the recommendation alternative. First, to meet quantitative requirements, the probabilistic linguistic term set is introduced to demonstrate the massive number of comments crawled. Second, considering preference and fluctuation, the relative importance of multiple attributes is determined. Because multiple attributes typically have cooperative or mutually exclusive relationships, a novel model is presented by introducing such relationship to modify relative importance. Third, inspired by the 2-additive Choquet integral operator and the Mahalanobis-Taguchi System, a bi-objective optimization model is proposed to illustrate the interactive effect of comments and develop an attribute correlation network. The specific relationships between attributes are reflected, including the positive and negative interactions. The relative importance, interactive imporantce and subgroup utility can be obtained. Fourth, to guarantee the operability and interpretability of the recommendation results, this paper presents a new information fusion operator and an probabilistic linguistic three-way recommendation process. Finally, a case study is used to demonstrate the complete procedures, and the parameter and comparative analyses highlight the effectiveness of the new operator and recommendation method.

Improved building MEP systems semantic segmentation in point clouds using a novel multi-class dataset and local–global vector transformer network

Point cloud semantic segmentation for mechanical, electrical, and plumbing (MEP) systems is crucial for establishing MEP systems digital twins. Deep learning has shown promise in automatically deriving point-wise semantic labels from point clouds. However, the utilization of 3D deep learning methods for directly segmenting complex, occluded, and densely distributed multi-class MEP systems in common building point clouds remains unexplored due to the lack of specialized public datasets. To address this, our study introduces a public, richly annotated, and multi-class point cloud dataset for MEP systems in common buildings, consisting of 92.10 million points across four buildings, 56 areas, and nine component categories. Subsequently, we propose Trans2Net, a novel and customized 3D deep learning method for the automatic and precise segmentation of MEP components, featuring local vector transformer module, global-aid vector transformer module, and learnable feature fusion operator that enhance the model’s sensitivity to local features, capture global context with sparse points, and dynamically fuse weighted local and global features. Experimental results demonstrate that Trans2Net significantly outperforms seven other state-of-the-art methods on the MEP dataset, achieving an overall accuracy of 97.39%, a mean accuracy of 95.45%, and a mean intersection over union (mIoU) of 90.76%. Notably, it surpasses the second-ranked method, PointMeta, by a margin of 9.91% in mIoU. This study provides a novel paradigm for automatically understanding MEP systems in point clouds and offers robust support for digital twin management of MEP systems, contributing to the prolonged lifespans of MEP systems.

PointSurFace: Discriminative point cloud surface feature extraction for 3D face recognition

Due to the geometric information in the 3D face data, the 3D face recognition methods exhibit better robustness against the physical attacks compared to the 2D recognition methods. In this paper, we propose PointSurFace, a 3D face point cloud recognition method integrated with a specially designed facial surface feature extraction (SurF) module. In PointSurFace, SurF comprehensively combines the coordinate positions, normal vectors, and angle relationships, etc., of each patch in the 3D face point cloud, to explore the explicit local geometric information of 3D faces and generate facial surface features. Subsequently, the surface features are fed into an encoder consisting of four set abstraction (SA) modules and three Inverted Residual MLP (InvResMLP) modules to extract more discriminative 3D facial features, where InvResMLP could suppress the overfitting and reduce information loss in the model training. In addition, PointSurFace utilizes channel fusion operations to integrate position information into the extracted 3D facial features, further enhancing the model performance. Experimental results in 3D facial recognition and verification across multiple datasets demonstrate that PointSurFace achieves the best 3D face recognition performance to date. For example, on Lock3DFace, it achieves the state-of-the-art recognition and verification accuracy of 90.03% and 80.01%, respectively, surpassing the previous methods by 2.02% and 3.19%. The ablation studies evaluate each module of PointSurFace, suggesting that the proposed surface feature extraction module significantly enhances the model performance.

A Two-Part Transformer Network for Controllable Motion Synthesis.

Although part-based motion synthesis networks have been investigated to reduce the complexity of modeling heterogeneous human motions, their computational cost remains prohibitive in interactive applications. To this end, we propose a novel two-part transformer network that aims to achieve high-quality, controllable motion synthesis results in real-time. Our network separates the skeleton into the upper and lower body parts, reducing the expensive cross-part fusion operations, and models the motions of each part separately through two streams of auto-regressive modules formed by multi-head attention layers. However, such a design might not sufficiently capture the correlations between the parts. We thus intentionally let the two parts share the features of the root joint and design a consistency loss to penalize the difference in the estimated root features and motions by these two auto-regressive modules, significantly improving the quality of synthesized motions. After training on our motion dataset, our network can synthesize a wide range of heterogeneous motions, like cartwheels and twists. Experimental and user study results demonstrate that our network is superior to state-of-the-art human motion synthesis networks in the quality of generated motions.

Research on the Collision Risk of Fusion Operation of Manned Aircraft and Unmanned Aircraft at Zigong Airport.

Low-altitude airspace is developing rapidly, but the utilization rate of airspace resources is low. Therefore, in order to solve the problem of the safe operation of the fusion of large UAVs and manned aircraft in the same airspace, this paper analyzes the theoretical calculation of the collision risk of the fusion operation of manned aircraft and UAVs at Feng Ming Airport in Zigong, verifying that while assessing the safety spacing of 10 km in the lateral direction, it further simulates the possibility of calculating the theoretical smaller safety spacing. The study will propose a new theory of error spacing safety margin and improve it according to the traditional Event collision risk model, combining the error spacing safety margin to establish an improved collision model more suitable for the fusion operation of manned and unmanned aircraft and reduce the redundancy of calculation. The error factors affecting manned and unmanned aircraft at Zigong Airport are analyzed, and theoretical calculations are analyzed by combining the actual data of Zigong Airport. Finally, the Monte Carlo simulation method is used to solve the error, substitute the calculation results, and simulate a section of the trajectory of the fusion operation for the reverse argument. The theoretical calculation results show that the collision risk from 10 km to 8 km satisfies the lateral target safety level (TSL) specified by ICAO under both traditional and improved models. The collision risk calculated by the improved model incorporating the error spacing safety margin is smaller, which enhances the safety of the model calculations. The results of the study can provide theoretical references for the fusion operation of manned and unmanned aircraft.

Prediction of protein secondary structure by the improved TCN-BiLSTM-MHA model with knowledge distillation

Secondary structure prediction is a key step in understanding protein function and biological properties and is highly important in the fields of new drug development, disease treatment, bioengineering, etc. Accurately predicting the secondary structure of proteins helps to reveal how proteins are folded and how they function in cells. The application of deep learning models in protein structure prediction is particularly important because of their ability to process complex sequence information and extract meaningful patterns and features, thus significantly improving the accuracy and efficiency of prediction. In this study, a combined model integrating an improved temporal convolutional network (TCN), bidirectional long short-term memory (BiLSTM), and a multi-head attention (MHA) mechanism is proposed to enhance the accuracy of protein prediction in both eight-state and three-state structures. One-hot encoding features and word vector representations of physicochemical properties are incorporated. A significant emphasis is placed on knowledge distillation techniques utilizing the ProtT5 pretrained model, leading to performance improvements. The improved TCN, achieved through multiscale fusion and bidirectional operations, allows for better extraction of amino acid sequence features than traditional TCN models. The model demonstrated excellent prediction performance on multiple datasets. For the TS115, CB513 and PDB (2018–2020) datasets, the prediction accuracy of the eight-state structure of the six datasets in this paper reached 88.2%, 84.9%, and 95.3%, respectively, and the prediction accuracy of the three-state structure reached 91.3%, 90.3%, and 96.8%, respectively. This study not only improves the accuracy of protein secondary structure prediction but also provides an important tool for understanding protein structure and function, which is particularly applicable to resource-constrained contexts and provides a valuable tool for understanding protein structure and function.

ELA-Net: An Efficient Lightweight Attention Network for Skin Lesion Segmentation.

In clinical conditions limited by equipment, attaining lightweight skin lesion segmentation is pivotal as it facilitates the integration of the model into diverse medical devices, thereby enhancing operational efficiency. However, the lightweight design of the model may face accuracy degradation, especially when dealing with complex images such as skin lesion images with irregular regions, blurred boundaries, and oversized boundaries. To address these challenges, we propose an efficient lightweight attention network (ELANet) for the skin lesion segmentation task. In ELANet, two different attention mechanisms of the bilateral residual module (BRM) can achieve complementary information, which enhances the sensitivity to features in spatial and channel dimensions, respectively, and then multiple BRMs are stacked for efficient feature extraction of the input information. In addition, the network acquires global information and improves segmentation accuracy by putting feature maps of different scales through multi-scale attention fusion (MAF) operations. Finally, we evaluate the performance of ELANet on three publicly available datasets, ISIC2016, ISIC2017, and ISIC2018, and the experimental results show that our algorithm can achieve 89.87%, 81.85%, and 82.87% of the mIoU on the three datasets with a parametric of 0.459 M, which is an excellent balance between accuracy and lightness and is superior to many existing segmentation methods.

Research on optimizing the controller configuration for high-power supply control system in nuclear fusion based on hybrid algorithm WOA-IAPSO

With the development of nuclear fusion technology, the high power supply control system plays a crucial role in the operation of nuclear fusion, and the PID controller is widely used in the field of power supply control. In this research, a novel hybrid algorithm is proposed that combines whale optimization algorithm (WOA) and the improved adaptive particle swarm optimization (IAPSO) to tune PID parameters in high-power supply. This organically combines the two methods into a new optimization strategy that is different from the simple weighted average method. By comparing the performance of WOA, IAPSO,SA,GA,SVM, and the hybrid WOA-IAPSO in step response under different working conditions, the simulation results show that the proposed algorithm can achieve more sensitive, stable and efficient response in high-power operation. This research is not only of great significance for the further development nuclear fusion technology, but also provides a useful reference for the application of adaptive optimization methods in other fields.

Entanglement and pseudo entanglement dynamics versus fusion in CFT

The fusion rules and operator product expansion (OPE) serve as crucial tools in the study of operator algebras within conformal field theory (CFT). Building upon the vision of using entanglement to explore the connections between fusion coefficients and OPE coefficients, we employ the replica method and Schmidt decomposition method to investigate the time evolution of entanglement entropy (EE) and pseudo entropy (PE) for linear combinations of operators in rational conformal field theory (RCFT). We obtain a formula that links fusion coefficients, quantum dimensions, and OPE coefficients. We also identify two definition schemes for linear combination operators. Under one scheme, the EE captures information solely for the heaviest operators, while the PE retains information for all operators, reflecting the phenomenon of pseudo entropy amplification. Irrespective of the scheme employed, the EE demonstrates a step-like evolution, illustrating the effectiveness of the quasiparticle propagation picture for the general superposition of locally excited states in RCFT. From the perspective of quasiparticle propagation, we observe spontaneous block-diagonalization of the reduced density matrix of a subsystem when quasiparticles enter the subsystem.

A study of the application of screw-in interlocking intramedullary needle in revision operation of tibial talar and fusion in one case.

A study of the application of screw-in interlocking intramedullary needle in revision operation of tibial talar and fusion in one case.

Biomechanical analysis of adjacent segments after correction surgery for adult idiopathic scoliosis: a finite element analysis

The biomechanical aspects of adjacent segment degeneration after Adult Idiopathic Scoliosis (AdIS) corrective surgery involving postoperative changes in motion and stress of adjacent segments have yet to be investigated. The objective of this study was to evaluate the biomechanical effects of corrective surgery on adjacent segments in adult idiopathic scoliosis by finite element analysis. Based on computed tomography data of the consecutive spine from T1–S1 of a 28-year-old male patient with adult idiopathic scoliosis, a three-dimensional finite element model was established to simulate the biomechanics. Two posterior long-segment fixation and fusion operations were designed: Strategy A, pedicle screws implanted in all segments of both sides, and Strategy B, alternate screws instrumentation on both sides. The range of motion (ROM), Maximum von Mises stress value of intervertebral disc (IVD), and Maximum von Mises stress of the facet joint (FJ) at the fixation adjacent segment were calculated and compared with data of the preoperative AdIS model. Corrective surgery decreased the IVD on the adjacent segments, increased the FJ on the adjacent segments, and decreased the ROM of the adjacent segments. A greater decrease of Maximum von Mises stress was observed on the distal adjacent segment compared with the proximal adjacent segment. The decrease of Maximum von Mises stress and increment of Maximum von Mises stress on adjacent FJ in strategy B was greater than that in strategy A. Under the six operation modes, the change of the Maximum von Mises stress on the adjacent IVD and FJ was significant. The decrease in ROM in the proximal adjacent segment was greater than that of the distal adjacent segment, and the decrease of ROM in strategy A was greater than that in strategy B. This study clarified the biomechanical characteristics of adjacent segments after AdIS corrective surgery, and further biomechanical analysis of two different posterior pedicle screw placement schemes by finite element method. Our study provides a theoretical basis for the pathogenesis, prevention, and treatment of adjacent segment degeneration after corrective surgery for AdIS.

Integrated energy cyber–physical system fusion modeling and operation state analysis under the cooperative attack

With the increasingly close coupling between the cyber system and the physical system, cyber attacks have a significant impact on the physical system by attacking the cyber system. As a cyber–physical system with multiple energy sources, integrated energy cyber–physical system (IECPS) is facing the risk of cyber attacks. However, the research on IECPS is still in the direction of operation optimization at the physical level, which cannot meet the requirements of system security and stability analysis under cyber cooperative attacks. Firstly, this paper analyzes the interaction mechanism between the energy network composed of electricity, heat and gas and the cyber network in IECPS, and put forward a layered modeling method of IECPS to analyze the influence of abnormal information flow caused by cyber attacks on energy flow. Secondly, based on the energy circuit method (ECM), a hybrid calculation method of IECPS energy-information flow is proposed to solve the model. Then, the load shedding state of IECPS and the tampering process of system control commands after the fake data injection attack are analyzed by using the proposed calculation method. Finally, the simulation analysis is carried out in an integrated energy system composed of IEEE 39-node power grid, 6-node heating network and 7-node gas network. The results show that, compared with the traditional Newton–Raphson method, the proposed IECPS hybrid calculation method can better obtain the change results of energy flow in the steady-state scenario and the cyber cooperative attack scenario, and identify the high-risk branches in the system, which is of great significance to the safe and stable operation of the energy system.