Double Transformer Research Articles

DBCvT: Double Branch Convolutional Transformer for Medical Image Classification

Convolutional Neural Networks (CNNs) are extensively utilized in medical disease diagnosis, demonstrating the prominent performance in most cases. However, medical image processing based on deep learning faces some challenges. The limited availability and time-consuming annotations of medical image data restrict the scale and accuracy of model training. Data diversity and complexity further complicate these challenges. In order to address these issues, we introduce the Double Branch Convolutional Transformer (DBCvT), a hybrid CNN-Transformer feature extractor, which can better capture diverse fine-grained features and remain suitable for small datasets. In this model, separable downsampling convolution (SDConv) is used to mitigate excessive information loss during downsampling in standard convolutions. Additionally, we propose the Dual branch Channel Efficient multi-head Self-Attention (DCESA) mechanism to enhance the self-attention efficiency, consequently elevating network performance and effectiveness. Moreover, we introduce a novel convolutional channel-enhanced Attention mechanism to strengthen inter-channel relationships within feature maps post self-attention. The experiments of DBCvT on various medical image datasets have demonstrated the outstanding classification performance and generalization capability of the proposed model.

Development and clinical validation of a deep learning-based knee CT image segmentation method for robotic-assisted total knee arthroplasty.

This study aimed to develop a novel deep convolutional neural network called Dual-path Double Attention Transformer (DDA-Transformer) designed to achieve precise and fast knee joint CT image segmentation and to validate it in robotic-assisted total knee arthroplasty (TKA). The femoral, tibial, patellar, and fibular segmentation performance and speed were evaluated and the accuracy of component sizing, bone resection and alignment of the robotic-assisted TKA system constructed using this deep learning network was clinically validated. Overall, DDA-Transformer outperformed six other networks in terms of the Dice coefficient, intersection over union, average surface distance, and Hausdorff distance. DDA-Transformer exhibited significantly faster segmentation speeds than nnUnet, TransUnet and 3D-Unet (p<0.01). Furthermore, the robotic-assisted TKA system outperforms the manual group in surgical accuracy. DDA-Transformer exhibited significantly improved accuracy and robustness in knee joint segmentation, and this convenient and stable knee joint CT image segmentation network significantly improved the accuracy of the TKA procedure.

Multi-sensor data fusion-enabled lightweight convolutional double regularization contrast transformer for aerospace bearing small samples fault diagnosis

Aiming at the problems of low information utilization and lack of feature mining capability in multi-sensor fusion networks, this study presents a multi-sensor data fusion-enabled lightweight convolutional double regularization contrast transformer for aerospace bearing small samples fault diagnosis. Firstly, a metric termed integrated cliff entropy is devised to assign weights to vibration signals from diverse sensor channels. It aims to enhance the cyclic impulse characteristics within the fused signals, thereby facilitating more precise fault identification. Secondly, a lightweight Diwaveformer architecture is constructed as the backbone of contrast learning. It enables the global and local features of faulty signals to be comprehensively extracted with less computational effort. Finally, a double contrast loss is constructed to optimize the distribution of intra-class and inter-class features to improve the fault identification ability of the network with small samples. Additionally, a discard regularization method is designed to remove the projection head during the contrast learning process, further advancing the model lightweight. Our method achieved accuracies of 95.54% and 92.56% on two aerospace bearing datasets with extremely sparse training samples, which proved its superior performance.

Thermodynamic analysis of an ammonia-water double absorption-resorption heat transformer system

Absorption heat transformer (AHT) presents an effective solution for enhancing energy efficiency through the utilization of low-grade industrial waste heat. To reduce the high operating pressures of traditional ammonia AHT and elevate waste heat to a higher temperature level, a novel ammonia-water double absorption-resorption heat transformer (DARHT) incorporating a coupled middle absorber-generator is proposed in this paper. A thermodynamic simulation model of the system is established using MATLAB software to analyze the effects of operating pressures and temperatures on the COP, temperature lift, solution circulation ratio and bypass ratio, thereby determining the optimal operating conditions of the system. The simulation results indicate that there exist optimal PH/PL pairs and middle absorber-generator temperatures which lead to the maximum COP of the system. COP decreases with the increase of absorber and resorber temperatures while increases with the increase of generator and desorber temperatures. When PH/PL pair is 1.5/0.36 MPa, the DARHT system can provide a temperature lift exceeding 70 °C, while the working pressures are significantly lower compared to conventional AHT, thereby reducing the requirements of robust components and enhancing the operational stability of the system.

Toroidal seawater conductivity sensors and instrumentation for anti-submarine warfare

This paper describes the development of toroid-based seawater conductivity sensors and instrumentation for 0 to 60 mS/cm conductivity range and 50 bar hydrostatic pressure operational capability which is equivalent to 500 metres of ocean depth. Double core toroidal transformer concept was used and investigated for Mn-Zn and Fe based nanocrystalline cores, turns ratio and temperature dependence. Different excitation and sense coil turn numbers were fabricated by keeping the turn&#x2019;s ratio 1:2, 1:4, 1:6 and 1:8. Optimum frequency of Mn-Zn based conductivity sensors with lower and higher turn numbers were observed for turn ratio 1:8 and 1:4 respectively at 120 kHz and obtained voltage responses of 952 and 1567 mV. Fe based Nano crystalline core with turn ratio 1:8 shows an optimum frequency of 70 kHz and giving a voltage difference of 830 mV. Two dimensional polynomials of order 2 were presented for temperature dependence of conductivity for Mn-Zn core based sensors. Sensor responses of Fe based nanocrystalline core sensors in &#x2018;lower turn number&#x2019; configuration with same turn ratio is presented and observed linear temperature dependence compared to Mn-Zn based sensors. The mechanical design and frequency of operation of developed conductivity sensors gives inherent immunity to ASW band of frequency interference. The efficacy of conductivity sensor with measuring electronics and commercial standard sensors profiling was proved experimentally compared up to 500 metres in sea profiling and obtained a root mean square error of 0.42 for profile data.

Analyzing Public Sentiment on the Amazon Website: A GSK-Based Double Path Transformer Network Approach for Sentiment Analysis

Analyzing Public Sentiment on the Amazon Website: A GSK-Based Double Path Transformer Network Approach for Sentiment Analysis

Adaptive interactive network for RGB-T salient object detection with double mapping transformer

Adaptive interactive network for RGB-T salient object detection with double mapping transformer

Advanced exergy analysis of a double absorption heat transformer recovering industrial waste heat for pure water production

Conventional and advanced exergy analyses are applied for a double absorption heat transformer system utilizing low-grade industrial waste heat for water purification. The conventional method provides information about the location and magnitude of inefficiencies and does not consider any interaction among components regarding the irreversibility occurring in them. The advanced exergy technique however, exposes the real origins of irreversibilities and latent potential for enhancement of each system component, along with reflecting mutual interaction between components. The simulation of the cycle is carried out with engineering equation solver software. The results show that the overall improvement potential of the cycle is only 16%. Also, the mutual interdependency between components is found to be weak as 97% of the total exergy destruction appeared as endogenous. Furthermore, considering endogenous avoidable exergy destruction, the improvement strategy of components obtained by advanced exergy method is disparate from conventional one. Results of the advanced exergy suggest this order: condenser, absorber, evaporator, absorber/evaporator assembly, generator assembly, economizer 1, and economizer 2. Additionally, a parametric investigation is performed to assess the sensitivity of splitting exergy destruction to the variation of heat source temperature for highlighting improvement opportunities. The results indicate that the condenser yields the highest amount of avoidable endogenous exergy destruction in the entire studied temperature range.

A double transformer residual super-resolution network for cross-resolution person re-identification

Cross-resolution person re-identification is a challenging problem in the field of person re-identification. In order to solve the problem of resolution mismatch, many studies introduce super-resolution into person re-identification tasks. In this work, we propose a cross-resolution person re-identification method based on double transformer residual super-resolution network (DTRSR), which mainly includes super-resolution module and person re-identification module. In the super-resolution module, we propose the double transformer network as our attention module. First of all, we divide the features extracted by the residual network. Then calculate the similarity between each local feature and the global feature after average pooling and maximum pooling respectively, which makes our module quickly capture the hidden weight information in the spatial domain. In the person re-identification module, we propose an effective fusion method based on key point features (KPFF). The key point extraction model can not only solve the problem that local features can not be accurately aligned, but also remove the interference of background noise. In order to fully mine the relationship between the features of each key point, we calculate the two-way correlation between each key point feature and other features, and then superimpose the two-way correlation with the feature itself to get the superposition feature which contains global and local information. The effectiveness of this method is proved by extensive experiments. Compared with the most advanced methods, the test results in the three datasets show that our method improves rank-1 by 1.1%, 3.5% and 1.7%; and rank-5 by 1.3%, 1.7% and 0.3%; and rank-10 by 0.1%, 0.4% and 0.1%, respectively.

Thermodynamic appraisal of utilizing compressor in double effect absorption heat transformer

In the current appraisal, three various designs of employing compressors in double-effect heat transformers are presented and evaluated from exergy and energy standpoints and collated with the simple configuration of mentioned designs. Water-LiBr is utilized as an operating fluid in noticed systems. As a function of the unit temperatures, the gross temperature lift, pressure ratio (R),and the efficiency of the economizer, a numerical model is also introducedto prognosticatethe irreversibility, exergy coefficient of performance (ECOP),coefficient of performance (COP), and the exergy destruction in each and every one of the key components (EF). The findings show that the generator has the highest exergy destruction, accounting for around 46.06% of the total exergy destruction. Additionally, a decrease in the temperature of the absorber increases the ECOP and COP. The outcomes indicate that systems utilizing water-LiBr as operating fluid have a higher outlet temperature of the absorber as well as higher condenser temperature and lower coefficient of performance (COP). the results showed that, Regarding the simple DEAHT curve, the COP slowly diminishes at low absorber-evaporator temperatures and rapidly at absorber temperatures more than 144.5 °C. Also, Utilizing the compressor in the basic design of the DEAHT unit results in improving the absorber temperature.

Control Strategy Based on Artificial Intelligence for a Double-Stage Absorption Heat Transformer

Thermal energy recovery systems have different candidates to mitigate CO2 emissions as recommended by the UN in its list of SDGs. One of these promising systems is thermal absorption transformers, which generally use lithium-water bromide as the working fluid. A Double Stage Heat Transformer (DSHT) is a thermal machine that allows the recovery of thermal energy at a higher temperature than it is supplied through the effect of steam absorption in a concentrated solution of lithium bromide. There are very precise thermodynamic models which allow us to calculate all the possible operating conditions of the DSHT. To perform the control of these systems, the use of Artificial Intelligence (AI) is proposed with two computational techniques—Fuzzy Logic (FL) and Artificial Neural Network (ANN)—to calculate in real-time the set of variables that maximize the product’s Gross Temperature Lift (GTL) and Coefficient of Performance (COP) in a DSHT. The values for Coefficient of Determination (R2), Mean Square Error Root (MRSE), and Mean Error Bias (MBE) for the two types of computational techniques were analyzed and compared with the purpose of identifying which of them may be more accurate to calculate the operating conditions (temperatures, pressures, concentration and flows) with the highest COP for an interval of the value of the temperature absorption entered by the user. The result of the analysis of the evaluated techniques concluded that the control strategy of a DSHT in real-time will be based on the precise calculation of the refrigerant flow in the second evaporator with a Neural Network of 30 neurons, 300 weights and 40 bias, as it is more accurate than the Fuzzy Logic technique. The goodness-of-fit for two computational techniques was evaluated as having an R2 higher than 0.98 for the provided data. Future AI controllers must be based on evaporator flow values with evaporator power at 3.9−04 kg/KJ.

Application of double core symmetrical phase shifting transformer in power flow regulation of UHV power grid

With the development of society, the demand for electric energy is growing day by day. However, the energy centers and load centers are geographically distributed in a reverse direction, so the power supply and demand are unbalanced. High voltage transmission lines are characterized by long distance and large capacity, and the power flow problems are increasingly prominent. In order to flexibly adjust the line power flow, improve the comprehensive utilization of grid assets, and improve the transmission capacity of the system, this paper analyzes the principle of double core symmetrical phase-shift transformer (DS-PST) and the application of power flow regulation. Firstly, the basic principle of phase shifting transformer is analyzed in detail, and the equivalent mathematical model of DS-PST and the calculation of main parameters are derived. Impedance coefficient and excitation coefficient are introduced to modify the parameter design of DS-PST. Then, an electromagnetic simulation model is established to verify the effectiveness of the design method. Finally, a system simulation model is built to verify the power flow regulation effect of DS-PST by taking 500 kV UHV power grid and IEEE standard multi-node system as examples. DS-PST can transfer the active power transmitted by the lines with high load rate to the lines with low load rate, which makes the power flow distribution of the system more balanced and effectively improves the transmission capacity of the power system.

Study on Vibration Characteristics of Transformer Winding

Compared with the existing winding deformation detection technologies, such as the frequency response method and short-circuit reactance, the vibration analysis method can effectively detect the loose state of windings. This paper has carried out mechanism analysis and experimental study on the change of winding modal characteristics caused by two kinds of defects, analyzed the transmission characteristics of three-phase double winding transformer winding vibration in the transformer oil and structure, studied the influence of pad falling off, plastic deformation and other defects on the mechanical state of the winding, calculated the winding short-circuit vibration signals corresponding to different mechanical states, and extracted the vibration characteristics that can effectively reflect the mechanical state of the winding. The research conclusion can provide an important basis for transformer fault diagnosis.

Research on fast solution method for vibration prediction of digital twin substation equipment based on neural network model

This study began by building a series of high-quality simulation modules of real space clones to obtain comparable data and conditions, and comparing them with data and situations obtained from the real state module to bring the system closer to reality. And the imaging model uses a number of numbers to generate the number of samples to train the neural network model of the plan. Finally, this paper takes the three-phase three column double winding transformer models of the module as an example to evaluate the proposed method of evaluating the design characteristics of the module.

Air Quality Classification and Measurement Based on Double Output Vision Transformer

The atmospheric quality monitor is an essential component of atmospheric protection. However, using complicated large-scale chemical equipment required by traditional methods is difficult for ordinary people. We are committed to allowing ordinary people to obtain a local air quality index (AQI) and then participate in the air quality monitoring of the entire district quickly and easily. So, we propose a method improved from Transformer in this article. This method processes pictures taken by mobile devices and predicts a more accurate local AQI level. The proposed method performs better than traditional methods in frequency and flexibility. Tokens in the Double Output Vision Transformer (DOViT) proposed in this article will be processed with the multihead self-attention (MSA) mechanism of the model to automatically extract the features in pictures for achieving higher classification accuracy. The application of Transformer makes our model more concise and efficient than previous CNN-based methods.

Intelligent Assistant Diagnosis System of Osteosarcoma MRI Image Based on Transformer and Convolution in Developing Countries.

Osteosarcoma is a malignant bone tumor commonly found in adolescents or children, with high incidence and poor prognosis. Magnetic resonance imaging (MRI), which is the more common diagnostic method for osteosarcoma, has a very large number of output images with sparse valid data and may not be easily observed due to brightness and contrast problems, which in turn makes manual diagnosis of osteosarcoma MRI images difficult and increases the rate of misdiagnosis. Current image segmentation models for osteosarcoma mostly focus on convolution, whose segmentation performance is limited due to the neglect of global features. In this paper, we propose an intelligent assisted diagnosis system for osteosarcoma, which can reduce the burden of doctors in diagnosing osteosarcoma from three aspects. First, we construct a classification-image enhancement module consisting of resnet18 and DeepUPE to remove redundant images and improve image clarity, which can facilitate doctors' observation. Then, we experimentally compare the performance of serial, parallel, and hybrid fusion transformer and convolution, and propose a Double U-shaped visual transformer with convolution (DUconViT) for automatic segmentation of osteosarcoma to assist doctors' diagnosis. This experiment utilizes more than 80,000 osteosarcoma MRI images from three hospitals in China. The results show that DUconViT can better segment osteosarcoma with DSC 2.6% and 1.8% higher than Unet and Unet++, respectively. Finally, we propose the pixel point quantification method to calculate the area of osteosarcoma, which provides more reference basis for doctors' diagnosis.

High-Frequency Current Transformers Cascade for Power Electronics Measurements.

High-frequency current transformers (HFCT) are widely used to measure fast transient current. Their advantages are simple structure and relatively moderate price. Their lower and upper bandwidth are limited, but the HFCT can be easily applied to many measuring applications in power electronics. The disadvantages of HFCT are substantial dimensions and a large weight. The paper proposes a system of a cascade connection of two transformers, which allows us to reduce these disadvantages. The properties of such an HFCT combination were investigated and described. In the article, the expression for double current transformer transmittances is derived. The frequency response of the sensor was determined, and the results were verified in a practical arrangement. An experimental setup of a cascade CT connection was made and tested, allowing for fast-changing signals in transients to be measured. This paper presents the theoretical basis and results of laboratory work on a wide range of static and dynamic tests of the proposed sensor.

Double Linear Transformer for Background Music Generation from Videos

Many music generation research works have achieved effective performance, while rarely combining music with given videos. We propose a model with two linear Transformers to generate background music according to a given video. To enhance the melodic quality of the generated music, we firstly input note-related and rhythm-related music features separately into each Transformer network. In particular, we pay attention to the connection and the independence of music features. Then, in order to generate the music that matches the given video, the current state-of-the-art cross-modal inference method is set up to establish the relationship between visual mode and sound mode. Subjective and objective experiment indicate that the generated background music matches the video well and is also melodious.

Primary Double Winding Coupling DC Transformer for DC Micro-Grid Applications

Primary Double Winding Coupling DC Transformer for DC Micro-Grid Applications

Energy grade splitting of hot water via a double effect absorption heat transformer

Energy level mismatch between waste heat and energy users poses huge barrier for local waste heat recovery. To tackle this challenge, an energy grade splitting strategy of hot water via absorption heat transformer is proposed. In this strategy, the hot water is separated into two branches. An absorption heat transformer extracts waste heat from the colder branch, and delivers heat output to increase the temperature of hotter branch. This process could deliver high temperature output and decrease the temperature of waste heat simultaneously, which meets the local waste heat recovery demand. Both experimental and modeling researches are carried out to investigate this strategy. Firstly, double effect LiBr-water absorption heat transformer is adopted for performance enhancement, and experimental investigation was carried out. Testing results showed that 391.4 kW heat output and COP of 0.618 were achieved, under colder branch temperature of 149.7 → 128.1 °C and hotter branch temperature of 151.0 → 128.1 °C. Secondly, modeling framework was built and used to evaluate the energy grade splitting performance under different conditions. Results showed that the proposed energy grade splitting strategy is effective under waste heat source temperatures of 130.0–160.0 °C, temperature lift of 5.0–20.0 °C, mass split ratio of 0.10–1.60 and temperature split ratio of 2.00–20.00. Such an effective and flexible strategy could further be promoted to other scenarios with different heat conversion technologies, and contribute to the wide application of local waste heat recovery.