Domain shift is a big challenge when deploying deep learning models in real-world applications due to various data distributions. The recent advances of domain adaptation mainly come from explicitly learning domain invariant features (e.g., by adversarial learning, metric learning and self-training). While they cannot be easily extended to multi-domains due to the diverse domain knowledge. In this paper, we present a novel multi-target domain adaptation (MTDA) algorithm, i.e., prompt-DA, through implicit feature adaptation for medical image segmentation. In particular, we build a feature transfer module by simply obtaining the domain-specific prompts and utilizing them to generate the domain-aware image features via a specially designed simple feature fusion module. Moreover, the proposed prompt-DA is compatible with the previous DA methods (e.g., adversarial learning based) and the performance can be continuously improved. The proposed method is evaluated on two challenging domain-shift datasets, i.e., the Iseg2019 (domain shift in infant MRI of different ages), and the BraTS2018 dataset (domain shift between high-grade and low-grade gliomas). Experimental results indicate our proposed method achieves state-of-the-art performance in both cases, and also demonstrates the effectiveness of the proposed prompt-DA. The experiments with adversarial learning DA show our proposed prompt-DA can go well with other DA methods. Our code is available at https://github.com/MurasakiLin/prompt-DA .

: The random failure of the unmanned aerial vehicles (UAVs) in performing the task of remote sensing will result in the lack of timely response in the mission execution process, affecting the overall result of the mission or causing partial or complete failure of a mission. To address the failure issues, a redundant and fault-tolerant method for UAV remote sensing networking were proposed. With this method, the lead-follower drone group flight mode is used. Multiple networked drones can complete the designated flight tasks stably and reliably. The simulation result verifies the efficacy of the redundant fault-tolerant method, which can help solve the problem of missing data or task failure in the remote sensing of UAV.

INTRODUCTION:Radiomics could be potential imaging biomarkers by capturing and analyzing the features. Children and adolescents with CHD have worse neurodevelopmental and functional outcomes compared with their peers. Early diagnosis and intervention are the necessity to improve neurological outcomes in CHD patients.METHODS:School-aged TOF patients and their healthy peers were recruited for MRI and neurodevelopmental assessment. LASSO regression was used for dimension reduction. ROC curve graph showed the performance of the model.RESULTS:Six related features were finally selected for modeling. The final model AUC was 0.750. The radiomics features can be potential significant predictors for neurodevelopmental diagnoses.CONCLUSION:The radiomics on the conventional MRI can help predict the neurodevelopment of school-aged children and provide parents with rehabilitation advice as early as possible.

The accelerating power of deep learning in diagnosing a disease and analyzing medical data will empower physicians and speed up decision-making in clinical environments. Applications of modern medical instruments and digitalization of medical care have generated large amounts of biomedical information in recent years. These pose challenges, demands, and opportunities for new AI methods and computational models for efficient data processing, analysis, and modeling with the generated data that are important for clinical applications and in understanding the underlying biological process.

In children with tetralogy of Fallot (TOF), there is a risk of brain injury even if intracardiac deformities are corrected. This population follow-up study aimed to identify the correlation between cerebral morphology changes and cognition in postoperative school-aged children with TOF. Resting-state functional magnetic resonance imaging (rs-fMRI) and the Wechsler Intelligence Scale for Children–Chinese revised edition (WISC-CR) were used to assess the difference between children with TOF and healthy children (HCs). Multiple linear regression showed that the TOF group had a lower verbal intelligence quotient (VIQ, 95.000 ± 13.433, p = 0.001) than the HC group and that VIQ had significant positive correlations with the cortical thickness of both the left precuneus (p < 0.05) and the right caudal middle frontal gyrus (p < 0.05) after adjustment for preoperative SpO2, preoperative systolic blood pressure (SBP), preoperative diastolic blood pressure (DBP) and time of aortic override (AO). Our results suggested that brain injury induced by TOF would exert lasting effects on cortical and cognitive development at least to school age. This study provides direct evidence of the relationship between cortical thickness and VIQ and of the need for strengthened verbal training in school-aged TOF patients after corrective surgery.

In the past, scholars used various computer vision and artificial intelligence methods to detect brain diseases via magnetic resonance imaging (MRI). In this paper, we proposed a novel system to detect sensorineural hearing loss (SNHL). First, we used three-level bior4.4 wavelet to decompose original brain image. Second, principal component analysis (PCA) was utilized for dimensionality reduction. Third, the generalized eigenvalue proximal support vector machine (GEPSVM) with Tikhonov regularization was employed as the classifier. The 10 repetitions of five-fold cross validation showed our method achieved an overall accuracy of 95.71 %. Our sensitivities over healthy control, left-sided SNHL, and right-sided SNHL are 96.00 %, 95.33 %, and 95.71 %, respectively. The proposed system is promising and effective in SNHL detection. It gives better performance than four state-of-the-art methods.

Alzheimer's disease (AD) is a progressive brain disease. The goal of this study is to provide a new computer-vision based technique to detect it in an efficient way. The brain-imaging data of 98 AD patients and 98 healthy controls was collected using data augmentation method. Then, convolutional neural network (CNN) was used, CNN is the most successful tool in deep learning. An 8-layer CNN was created with optimal structure obtained by experiences. Three activation functions (AFs): sigmoid, rectified linear unit (ReLU), and leaky ReLU. The three pooling-functions were also tested: average pooling, max pooling, and stochastic pooling. The numerical experiments demonstrated that leaky ReLU and max pooling gave the greatest result in terms of performance. It achieved a sensitivity of 97.96%, a specificity of 97.35%, and an accuracy of 97.65%, respectively. In addition, the proposed approach was compared with eight state-of-the-art approaches. The method increased the classification accuracy by approximately 5% compared to state-of-the-art methods.

Magnetic resonance (MR) imaging is widely used in daily medical treatment. It could help in pre-surgical, diagnosis, prognosis, and postsurgical processes. It could be beneficial for diagnosis to classify MR images of brain into healthy or abnormal automatically and accurately, since the information set MRIs generate is too large to interpret with manual methods. We propose a new approach with wavelet-entropy as the features and the kernel based extreme learning machine (K-ELM) to be the classifier. Our method employs 2D-discreet wavelet transform (DWT), and calculates the entropy as features. Then, a K-ELM is trained to classify images as pathological or healthy. A 10 × 10-fold cross validation is conducted to prevent overfitting. The method achieves the sensitivity as 97.48 %, the specificity as 94.44 %, and the overall accuracy as 97.04 % based on 125 MR images. The performance suggests the classifier is robust and effective by comparison with the recently published approaches.

AIM:Sensorineural hearing loss is correlated to massive neurological or psychiatric disease.MATERIALS:T1-weighted volumetric images were acquired from fourteen subjects with right-sided hearing loss (RHL), fifteen subjects with left-sided hearing loss (LHL), and twenty healthy controls (HC).METHOD:We treated a three-class classification problem: HC, LHL, and RHL. Stationary wavelet entropy was employed to extract global features from magnetic resonance images of each subject. Those stationary wavelet entropy features were used as input to a single-hidden layer feedforward neuralnetwork classifier.RESULTS:The 10 repetition results of 10-fold cross validation show that the accuracies of HC, LHL, and RHL are 96.94%, 97.14%, and 97.35%, respectively.CONCLUSION:Our developed system is promising and effective in detecting hearing loss.

(Aim) Unilateral sensorineural hearing loss is a brain disease, which causes slight morphology within brain structure. Traditional manual method can ignore this change. (Method) First, we used dual-tree complex wavelet transform to extract features. Afterwards, we used kernel principal component analysis to reduce feature dimensionalities. Finally, multinomial logistic regression was employed to be the classifier. (Result) The 10 times of 10-fold stratified cross validation showed our method achieved an overall accuracy of 96.17 ± 2.49 ± 2.58 ± 2.42 ± 3.16

Objective:: To evaluate the characteristics of atherosclerotic middle cerebral artery (MCA) stenosis by high-resolution magnetic resonance imaging (HR-MRI) and determine the relationship between wall characteristics and infarction patterns. Methods:: Thirty-six patients with acute ischaemic stroke due to MCA stenosis underwent diffusion-weighted magnetic resonance imaging (DWI) and HR MRI. Wall characteristics of MCA, including irregular surface, superior location, T2-hyperintense of plaques and positive remodelling (PR), were analysed. Characteristics of acute infarct on DWI were categorised according to the number (single or multiple infarcts) and the pattern of cerebral infarcts (cortical, border zone or perforating artery territory infarcts). The relationship between wall characteristics and infarction patterns was evaluated. Results:: PR was observed in 20 patients, irregular surface plaque in 18 patients, superior location of plaques in 14 patients and T2-hyperintense foci in 13 patients. Seventeen patients had multiple acute cerebral infarcts and 13 showed single acute cerebral infarcts. Border zone infarcts were the most common (76.5%) among multiple acute infarcts. Penetrating artery infarcts (PAI) accounted for 76.9% of all single infarcts. Multiple infarcts were more frequently observed in patients with PR (P =  0.007) or plaque surface irregularity (P =  0.035). Single infarcts, especially PAI, were more prevalent in patients with superior plaque (P =  0.030). No statistically significant differences were observed between multiple and single infarcts in patients with T2-hyperintense lesions (P = 0.638). Conclusions:: PR or irregular surface plaques were associated with artery-to-artery embolism. Superior location of plaques was associated with PAI. HR-MRI provides insights into intracranial atherosclerosis in vivo, predictive of infarction patterns.

Diffusion tensor imaging (DTI) offers an efficient evaluation over microstructure of white matter integrity. In this study, we aim to study the white matter tracts, which are associated with normal ageing. 42 healthy subjects were included, which is composed of two groups. One is young subjects (age = 22.29±1.35), and the other is elder (age = 63.71±2.00). Statistical analysis over FA was carried out by TBSS. We observed FA decreases significantly in several white matter tracts, including anterior thalamic radiation (ATR), cingulum (cingulate gyrus), corticospinal tract (CST), forceps minor, forceps major, inferior longitudinal fasciculus (ILF), superior longitudinal fasciculus (SLF). Most found all tracts were reported in past literatures. Nevertheless, we are the first to find the relationship between forceps major and ageing.

The technology of AS?i can solve the difficulty of wiring in industrial field. To extend the compatibility of the con?troller with AS?I,the characteristics of AS?i are described in this paper. The communication interface of AS?i was designed to rea?lize the function that the controller communicates with other slaves via AS?i. Based on GB/T18858.2?2012,PIC16F916 was cho?sen as CPU and AS?i4I?GE?MT as communication converting interface chip to realize the functions of the network communication of controller with AS?i. The testing results show that the interface designed in this paper can communicate with other AS?i mo?dules properly.

Alterations of brain structure and functional connectivity have been described in patients with hearing impairments due to distinct pathogenesis; however, the influence of unilateral hearing loss (UHL) on brain morphology and regional brain activity is still not completely understood. In this study, we aim to investigate regional brain structural and functional alterations in patients with UHL. T1-weighted volumetric images and task-free fMRIs were acquired from 14 patients with right-sided UHL (pure tone average ≥ 40 dB HL) and 19 healthy controls. Hearing ability was assessed by pure tone audiometry. Voxel-based morphometry (VBM) was performed to detect brain regions with changed gray matter volume or white matter volume in UHL. The amplitude of low-frequency fluctuation (ALFF) was calculated to analyze brain activity at the baseline and was compared between two groups. Compared with controls, UHL patients showed decreased gray matter volume in bilateral posterior cingulate gyrus and precuneus, left superior/middle/inferior temporal gyrus, and right parahippocampal gyrus and lingual gyrus. Meanwhile, patients showed significantly decreased ALFF in bilateral precuneus, left inferior parietal lobule, and right inferior frontal gyrus and insula and increased ALFF in right inferior and middle temporal gyrus. These findings suggest that chronic UHL could induce brain morphological changes and is associated with aberrant baseline brain activity.

A method for software automatic focusing is proposed to meet the requirements of the fast and the large number of focusing. By processing on the CCD image acquired and using the focusing algorithm based on combining image memory and the traditional evaluation functions presented in this paper,a series of points which have accurate focus positions on the scanning plane are obtained. In order to speed up the scanning speed and improve the efficiency of the system,the others on the plane complete focus on the basis of the focal plane built relying on the focused point. Considering the detection chip may not be an absolute plane,the focal plane is established by least square method. Focal plane meets the requirements of the equipment which is built by selecting 6 points for precise focus,proved by experiments.

The article determines the overall scheme according to the processing requirements.It uses CNC and servo system to build a semi-closed loop axis linkage control system in order to complete work piece in feed and indexing and grinding wheel interpolation.At the same time,it increases a wheel dresser axis for dressing grinding wheel profile in real time to ensure the quality.

The traditional RS 485 makes the efficiency of communication low and causes data confusion and loss.To optimize the RS 485 interface circuit for the purpose of increasing data transmission efficiency,a novel 485 repeater was designed in this paper.In this system,the hardware includes photoelectric isolation,electric control and zero delay switching circuits.On the other hand,by means of the simulation of serial ports which is realized by the software,the system makes up the insufficient caused by single serial port of AT89C51 and makes full use of the pins of MCU.The system has been applied in automatic meter reading system,which has the characteristics of good real-time performance,high anti-interference,low cost,stability and reliability.

WinCC(Windows Control Center) is the configuration software produced by SIEMENS,which integrates the HMI system and monitoring management system.Through the design of supervision control interfaced and animations,configuration and the use of VB script,the monitoring programming of the Modular Manufacturing System is designed based on the WinCC.The system runs smoothly,so we can conclude that the design is reasonable and the software is easy to use.

IEEE1394 bus is a kind of serial high-speed bus for the real-time transmission of data.Due to the advantage in bandwidth,speed and management,IEEE1394 bus has become the principal choice of the application of industrial automation and navigation.At the same time,DSP has also already taken very large market share,so how to link the IEEE1394 bus to DSP chips becomes very important.Proposed a method of IEEE1394 optical bus interface design based on DSP(TMS320C6747) according to the background.By using the DSP(TMS320C6747) and IEEE1394 chips,the paper does the work of hardware design and simulation.The results of simulation show that all critical signal's integrity of the circuit is good enough to guarantee stable operation of the hardware.The work provides the technonlgy to make the high-speed equipments based on DSP be able to link to the IEEE1394 optical bus net.All the work will accelerate the application and development of high-speed electronical system.