Dynamic stability refers to the ability of a magnetic actuation system to maintain equilibrium by damping oscillations. It is typically quantified by physical scalars that serve as critical indicators of both safety and effectiveness in real-world applications. Traditional real-time assessment methods often require measuring and calibrating task-specific parameters. While feasible in principle, these approaches are hindered in practice by procedural complexity, time consumption, measurement difficulty, and uncertainties inherent to the non-contact, non-rigid nature of magnetic actuation. In this work, we eliminate the need for parameter assumptions and direct measurements, enabling a more efficient and generalizable evaluation of dynamic stability. Our approach implicitly incorporates uncertain physical properties into a learning-based framework. Specifically, we propose a temporal difference policy that predicts dynamic stability by comparing multiple time-varying sequences, thereby reducing the influence of task-specific parameters. The robustness and effectiveness are validated through extensive experiments, including baseline comparisons and online implementations across diverse magnetically actuated objects. Our findings highlight its practical advantages and pave the way for innovative control strategies in precise magnetic actuation applications.

Endovascular intervention relies on accurate path planning and precise guidewire manipulation, which makes training difficult and time-consuming. We present a real-time vascular intervention simulator that couples Cosserat rod-based guidewire dynamics, mesh-contraction centerline extraction with shortest-path planning, and a Reinforcement Learning (RL) controller for autonomous navigation. A Proximal Policy Optimization (PPO) agent with a curriculum schedule and a visibility-aware error-correction strategy learns push-pull-rotate actions aligned with waypoint paths. The system runs at 30 Hz in Unreal Engine, achieves a success rate of over 95 % on integrated and real-model scenarios, and reduces the completion time compared with a distance-reward RL baseline and a rule-based policy. Results indicate that integrating physically-plausible simulation with path planning and curriculum-driven PPO agent yields robust recovery from deviations and efficient navigation.

Medical image translation is of great value but is very difficult due to the requirement with style change of noise pattern and anatomy invariance of image content. Various deep learning methods like the mainstream GAN, Transformer and Diffusion models have been developed to learn the multi-modal mapping to obtain the translated images, but the results from the generator are still far from being perfect for medical images. In this paper, we propose a robust multi-contrast translation framework for MRI medical images with knowledge distillation and adversarial attack, which can be integrated with any generator. The additional refinement network consists of teacher and student modules with similar structures but different inputs. Unlike the existing knowledge distillation works, our teacher module is designed as a registration network with more inputs to better learn the noise distribution well and further refine the translated results in the training stage. The knowledge is then well distilled to the student module to ensure that better translation results are generated. We also introduce an adversarial attack module before the generator. Such a black-box attacker can generate meaningful perturbations and adversarial examples throughout the training process. Our model has been tested on two public MRI medical image datasets considering different types and levels of perturbations, and each designed module is verified by the ablation study. The extensive experiments and comparison with SOTA methods have strongly demonstrated our model's superiority of refinement and robustness.

Arbitrary Style Transfer (AST) renders an image by adopting the style of any chosen artwork while preserving its content structure. Despite the widespread popularity of feedforward AST methods, they tend to merely optimize the statistical characteristics of images, leading to unnatural outputs and displeasing low-quality distortions. In contrast, diffusion models effectively address this issue by reconstructing the overall image. Unlike typical image generation tasks controlled by a single condition, style transfer demands the simultaneous consideration of multiple conditions. We propose Condition Harmony DiffusionNet (CHDNet), which distinguishes between style and content conditions, integrating them into harmonious conditions to collaboratively guide the generation process. We innovatively introduce a content aware attention, designed to extract semantic features of the content image across multiple dimensions, distinctly setting it apart from the style condition. Furthermore, we have improved the skip connections in the diffusion model, which introduces a slight increase in model complexity but results in a substantial improvement in the representation of fine details. Further refinements in the image sampling process empower us with great control over the stylization effect in the generated results. Our method successfully employs diffusion models via harmonious conditions to solve AST, achieving outstanding effects. Experiments demonstrate that our method achieves stateof-the-art arbitrary style transfer.

Medical image translation holds significant value, but its difficulty is amplified due to variations in noise patterns and the requisite anatomical invariance of image content. Various deep learning approaches, such as mainstream Generative Adversarial Networks (GANs), have been developed to learn multimodal mappings for obtaining translated images. However, the results produced by generators remain far from perfect for medical images, given the challenging requirements of style variations in noise patterns and anatomical invariance. In this paper, a medical image translation framework is proposed based on a diffusion model and knowledge distillation. To enhance the robustness of adversarial training and the accuracy of generated images, unlike traditional GANs, this framework incorporates an adaptive forward diffusion module for data augmentation following the generator. Additionally, the discriminator is designed as a timestep-dependent discriminator. Both real and generated images undergo the same forward diffusion process, and the discriminator learns to discriminate between real and generated images at each time step. Finally, an additional refinement network is composed of structurally similar but differently inputted teacher and student modules. Unlike existing knowledge distillation approaches, our teacher module is designed as a registration network with more inputs to better learn noise distribution and further refine translation results during training. Subsequently, knowledge is thoroughly distilled into the student module to ensure the generation of superior translation results. Extensive experiments on two public medical image datasets, along with comparisons with SOTA methods, demonstrate that the model produces higher quality and more robust images.

Human attention is one of the important indicators of human normal activity, and the classification of EEG signals helps determining the level of human attention. However, most existing EEG signal classification algorithms based on deep learning networks suffer from the difficulty in parallelization and low efficiency. In this paper, a new model of EEG signal classification is proposed including feature extraction, temporal feature encoding and classification modules. The feature extraction module adopts the two-branch multiscale CNN model and combines the G-bneck module, which improves the efficiency of the model by generating more important characteristics through the channel attention mechanism. In temporal feature encoding module, the Transformer encoder is utilized with parallel computing ability to extract temporal features, and the causal convolution-based position coding method is presented to help solve the translation homogeneity problem. Our algorithm has been tested with intrasubjects and inter-subjects on both self-built and public datasets, and the experimental results prove that the new model achieves competitive classification performance. Our model is used with a single-channel EEG signal acquisition device for human attention, and is also applicable to multi-channel EEG signals.

Internet of Medical Things (IoMT) technology has significantly helped surgeons perform complex clinical procedures, including interventional and endoscopic surgeries. However, surgeons face challenges in quickly becoming proficient with some IoMT devices. This difficulty stems from the fact that IoMT devices are not commonly used or familiar in their daily work and lives. To address this issue, we propose a novel interactive framework for IoMT, named freehand interaction, which includes visual control and haptic feedback. Our idea is to allow surgeons to operate surgical instruments, such as needles, capsules, and catheters with their bare hands and regular experience. At the same time, identical instruments in the surgical environment mirror the surgeon’s actions through visual control. The interactive forces encountered in the surgical environment are quantitatively communicated to the surgeon through the hand-held instruments. Our IoMT framework achieves both visual control and haptic feedback using electromagnetic mechanisms, ensuring mid-air freehand manipulation and contactless remote actuation. We perform different tasks in suspended, liquid, and in-vitro environments. The hand-held and mirroring instruments show high similarity and correlation, even within different electromagnetic systems and confined workspaces. Tracking accuracy, response time, and haptic forces quantified by a mechanical gauge are satisfactory. Our algorithm processes raw video streams and maintains efficiency even in scenarios with partial hand occlusion and various types of image noise. In summary, this work contributes to improving visual intuition and haptic immersion in IoMT applications.

Abstract With the development of virtual reality technology, simulation surgery has become a low‐risk surgical training method and high‐precision positioning of surgical instruments is required in virtual simulation surgery. In this paper we design and validate a novel electromagnetic positioning method based on a uniform gradient magnetic field. We employ Maxwell coils to generate the uniform gradient magnetic field and propose two positioning algorithms based on magnetic field, namely the linear equation positioning algorithm and the magnetic field fingerprint positioning algorithm. After validating the feasibility of proposed positioning system through simulation, we construct a prototype system and conduct practical experiments. The experimental results demonstrate that the positioning system exhibits excellent accuracy and speed in both simulation and real‐world applications. The positioning accuracy remains consistent and high, showing no significant variation with changes in the positions of surgical instruments.

Electroencephalogram (EEG) plays an important role in studying brain function and human cognitive performance, and the recognition of EEG signals is vital to develop an automatic sleep staging system. However, due to the complex nonstationary characteristics and the individual difference between subjects, how to obtain the effective signal features of the EEG for practical application is still a challenging task. In this article, we investigate the EEG feature learning problem and propose a novel temporal feature learning method based on amplitude-time dual-view fusion for automatic sleep staging. First, we explore the feature extraction ability of convolutional neural networks for the EEG signal from the perspective of interpretability and construct two new representation signals for the raw EEG from the views of amplitude and time. Then, we extract the amplitude-time signal features that reflect the transformation between different sleep stages from the obtained representation signals by using conventional 1-D CNNs. Furthermore, a hybrid dilation convolution module is used to learn the long-term temporal dependency features of EEG signals, which can overcome the shortcoming that the small-scale convolution kernel can only learn the local signal variation information. Finally, we conduct attention-based feature fusion for the learned dual-view signal features to further improve sleep staging performance. To evaluate the performance of the proposed method, we test 30-s-epoch EEG signal samples for healthy subjects and subjects with mild sleep disorders. The experimental results from the most commonly used datasets show that the proposed method has better sleep staging performance and has the potential for the development and application of an EEG-based automatic sleep staging system.

Micro-expressions are spontaneous, rapid and subtle facial movements that can neither be forged nor suppressed. They are very important nonverbal communication clues, but are transient and of low intensity thus difficult to recognize. Recently deep learning based methods have been developed for micro-expression (ME) recognition using feature extraction and fusion techniques, however, targeted feature learning and efficient feature fusion still lack further study according to the ME characteristics. To address these issues, we propose a novel framework Feature Representation Learning with adaptive Displacement Generation and Transformer fusion (FRL-DGT), in which a convolutional Displacement Generation Module (DGM) with self-supervised learning is used to extract dynamic features from onset/apex frames targeted to the subsequent ME recognition task, and a well-designed Transformer Fusion mechanism composed of three Transformer-based fusion modules (local, global fusions based on AU regions and full-face fusion) is applied to extract the multi-level informative features after DGM for the final ME prediction. The extensive experiments with solid leave-one-subject-out (LOSO) evaluation results have demonstrated the superiority of our proposed FRL-DGT to state-of-the-art methods.

To meet the need of 3-D positioning of surgical instrument in interventional surgery and the requirement of smaller size sensor for the narrow blood vessels, a new electromagnetic positioning model is proposed with a simplified receiver. Based on the electromagnetic theory and geometry principle, the electromagnetic field transmitter with groups of three orthogonal coils and the simplified receiver with a smaller size than existing sensors are designed. Then, using the Biot–Savart law, the distance between receiving end and geometric center of orthogonal coils is calculated, and the spatial coordinate of receiving end is computed with the spherical intersection formula. To further reduce the positioning error, a two-round accuracy improvement algorithm is designed for selecting the optimal topology of coil groups at the transmitter end and fitting correction of the calculated distance. We implement the prototype system and perform both simulation experiments and actual experiments. The results show that our proposed new electromagnetic positioning method with a simplified receiver has higher accuracy and stronger stability, compared with the traditional positioning approach of three-coil transmitter and three-coil receiver.

Performing automatic retinal vessel segmentation on fundus image can obtain clear retinal vessel structure quickly, which will assist doctors to improve the efficiency and reliability of diagnosis. In fundus image, there are many small vessels and some areas with low contrast, and there may be abnormal areas. Therefore, achieving automatic retinal vessel segmentation with high performance is still challenging. The retinal vessel in the image is a topological structure, so the distribution of retinal vessel pixels in each pixel row (or column) should have some relationship to other rows (or columns). Motivated by this observation, we propose Pixel Rows and Columns Relationship Modeling Network (PRCRM-Net) to achieve high-performance retinal vessel segmentation. PRCRM-Net separately models the relationship between different pixel rows and pixel columns of fundus image, and achieves retinal vessel segmentation by classifying the pixels in units of pixel row and pixel column. The input of PRCRM-Net is the feature map extracted by U-Net. PRCRM-Net firstly processes the input feature map into row feature sequence and column feature sequence respectively. Secondly, it models the relationship between the elements in the row feature sequence and column feature sequence respectively based on Transformer. Finally, the updated row feature sequence and column feature sequence are used to obtain row-based segmentation result and column-based segmentation result respectively. And the final segmentation result is the combination of these two types of results. To evaluate the performance of PRCRM-Net, we conduct comprehensive experiments on three representative datasets, DRIVE, STARE and CHASE_DB1. The experiment results show that the proposed PRCRM-Net achieves state-of-the-art performance.

Myocardial infarction (MI) can cause serious harm to the human body. For patients with acute MI, coronary intervention is the treatment of choice, and electrocardiogram (ECG) is a useful tool for diagnosing the type and location of MI. With the rapid development of the 5th Generation (5G) newtork technology, remote interventional surgery will be the future development trend. In this paper, a 12-lead ECG signal acquisition device that can be used for wireless communication is designed. We propose a multi-branch Mobi-Trans model for MI localization, taking each lead of the ECG signal as the input of each corresponding branch. For each branch, we first extract features using lightweight depthwise convolution and channel-based attention mechanism and then use self-attention with relative position representations in the improved Transformer module to make the model pay more attention to more important positions in the ECG signal, and finally we use the branch attention module to enable the model to focus more on the branches that contribute more to the localization. The experimental results on the PTB database show that our proposed model achieves an overall accuracy of 99.91%, so the Mobi-Trans model can be used in conjunction with our ECG signal acquisition device to assist in the localization of lesions in remote interventional cardiac surgery in the future, and perform real-time ECG monitoring during the surgery.

Object recognition with multimodal representation has recently attracted great interest in the field of intelligent robotics, where visual and tactile fusion learning shows the potential to improve performance. However, existing approaches primarily focus on capturing the complementary features from two modalities, while ignoring the disparities between vision and touch, and overlooking the fusion of features from different scales. In this article, we propose an alignment and multi-scale fusion method (AMSF) for robotic object recognition to address these challenges. The proposed method exploits a novel alignment strategy based on contrastive learning through multimodality information and provides a more grounded visual and tactile representation. In the fusion of extracting interactive information, a multi-scale fusion module by transformer is applied to integrate features on different scales from two modalities and generates an ideal representation for a pair of visual and tactile data. Abundant experiments are carried out on three public datasets, and the results validate the superiority of our method. Furthermore, the effectiveness of the proposed two modules has been illustrated in ablation studies.

For various remote sensing objects, the current detection framework based on a single detection pipeline fails to provide satisfactory detection accuracy. In order to further improve the object recognition ability of the detection model, this article introduces the effective “multiexpert” mechanism into the field of remote sensing object detection and then constructs a multiexpert detection network (MEDNet). In this model, we first construct multiple feature pyramids (MFPs) to replace the traditional single feature pyramid to enrich the semantic representation ability of the model. Then, we equip multiple detection experts (MDEs) to leverage multiple kinds of features from MFP to perform different semantic predictions. As the first CNN-based multiexpert detection model for remote sensing images, we tailor a loss distance-based k-experts clustering (LD-kEC) strategy to assign training samples to different detection experts in an unsupervised fashion. By this strategy, we can directly use the existing remote sensing dataset without expert labels for end-to-end training of our multiexpert model. The experimental results prove that the proposed multiexpert-based detector can indeed significantly improve the object detection performance for remote sensing images.

Limited by the GPU memory, the current mainstream detectors fail to directly apply to large-scale remote sensing images for object detection. Moreover, the scale range of objects in remote sensing images is much wider than that of general images, which also greatly hinders the existing methods to effectively detect geospatial objects of various scales. For achieving high-performance object detection on large-scale remote sensing images, this article proposes a much faster and more accurate detecting framework, called cropping region proposal network-based scale folding network (CRPN-SFNet). In our framework, the CRPN includes a weak semantic RPN for quickly locating interesting regions and a strategy of generating cropping regions to effectively filter out meaningless regions, which can greatly reduce the computation and storage burden. Meanwhile, the proposed SFNet leverages the scale folding-based training and testing methods to extend the valid detection range of existing detectors, which is beneficial for detecting remote sensing objects of various scales, including very small and very large geospatial objects. Extensive experiments on the public Dataset for Object deTection in Aerial images data set indicate that our CRPN can help our detector deal the larger image faster with the limited GPU memory; meanwhile, the SFNet is beneficial to achieve more accurate detection of geospatial objects with wide-scale range. For large-scale remote sensing images, the proposed detection framework outperforms the state-of-the-art object detection methods in terms of accuracy and speed.

Electrocardiogram(ECG) signal quality assessment(SQA) plays an essential role in significantly improving the ECG signal quality. However, the reliability and accuracy of SQA based on wearables in free-living conditions are extremely limited by various noises such as motion artifacts(MA), which have morphology similar to some of the ECG signals. We propose a novel ECG SQA method based on low-dimensional depth local dual-view(DLDV) features and the Transformer model. First, the Empirical Mode Decomposition (EMD) and Fast Fourier Transform (FFT) are combined to achieve the DLDV (the DLDV features can distinguish the nuances between the ECG and MA) feature extraction. Then, the proposed Transformer based model can embed DLDV features and the ECG signal into low-dimensional space and then use the embedded feature vectors to train the classifier, which classifies signal segments into “acceptable” or “unacceptable” categories to achieve SQA. Note that the low-dimensional embedding makes it easier and faster for Transformer to learn features that can distinguish the local nuances between the ECG and MA. Finally, we conduct extensive experiments on the Physionet Computing in Cardiology Challenge 2011 database using both the original and balanced datasets. Experimental results show that the accuracy achieves 99.82%. We also test the contribution of different features to quality classification, and the experiment proved that the DLDV features are the most improved classification accuracy(from 97.53% to 99.82%). Compared with the state-of-the-art methods, the proposed method achieves the best classification result and shows superior performance.

EEG plays an important role in the analysis and recognition of brain activity, and which has great potential in the field of biometrics, while EEG-based time series classification is complicated and difficult due to the nonstationary characteristics and individual difference. In this paper, we investigate the EEG signal classification problem and propose a multi-subband and multi-subepoch time series feature learning (MMTSFL) method for automatic sleep stage classification. Specifically, MMTSFL first decomposes multiple subbands with various frequency from raw EEG signals and partitions the obtained subbands in-to multiple consecutive subepochs, and then employs time series feature learning to obtain effective discriminant features. Moreover, amplitude-time based signal features are extracted from each subepoch to represent dynamic variation of EEG signals, and MMTSFL conduct further multipurpose feature learning for specific features, consistent features and temporal features simultaneously. Experiment results on three classification tasks of sleep quality evaluation, fatigue detection and sleep disease diagnosis demonstrate the superiority of the proposed method.

Sleep staging based on electroencephalogram (EEG) plays an important role in the clinical diagnosis and treatment of sleep disorders. In order to emancipate human experts from heavy labeling work, deep neural networks have been employed to formulate automated sleep staging systems recently. However, EEG signals lose considerable detailed information in network propagation, which affects the representation of deep features. To address this problem, we propose a new framework, called MRNet, for data-driven sleep staging by integrating a multi-scale feature fusion model and a Markov-based sequential correction algorithm. The backbone of MRNet is a residual block-based network, which performs as a feature extractor. Then the fusion model constructs a feature pyramid by concatenating the outputs from the different depths of the backbone, which can help the network better comprehend the signals in different scales. The Markov-based sequential correction algorithm is designed to reduce the output jitters generated by the classifier. The algorithm depends on a prior stage distribution associated with the sleep stage transition rule and the Markov chain. Experiment results demonstrate the competitive performance of our proposed approach on both accuracy and F1 score (e.g., 85.14% Acc and 78.91 % F1 score on Sleep-EDFx, and 87.59% Acc and 79.62% F1 score on Sleep-EDF). All codes are provided publicly online.11https://github.com/XueJiang16/MRNet

EEG signals are usually simple to obtain but expensive to label. Although supervised learning has been widely used in the field of EEG signal analysis, its generalization performance is limited by the amount of annotated data. Self-supervised learning (SSL), as a popular learning paradigm in computer vision (CV) and natural language processing (NLP), can employ unlabeled data to make up for the data shortage of supervised learning. In this paper, we propose a self-supervised contrastive learning method of EEG signals for sleep stage classification. During the training process, we set up a pretext task for the network in order to match the right transformation pairs generated from EEG signals. In this way, the network improves the representation ability by learning the general features of EEG signals. The robustness of the network also gets improved in dealing with diverse data, that is, extracting constant features from changing data. In detail, the network's performance depends on the choice of transformations and the amount of unlabeled data used in the training process of self-supervised learning. To verify the effectiveness of the proposed method, we take Sleep-edf, Sleep-edfx, Dod-O, and Dod-H datasets with more than 300k samples to conduct our experiments. Empirical evaluations on the Sleep-edf dataset demonstrate the competitive performance of our method on sleep staging (88.16% accuracy and 81.96% F1 score) and verify the effectiveness of SSL strategy for EEG signal analysis in limited labeled data regimes. All codes are provided publicly online.11https://githuh.com/XueJiang16/ssl-torch

Convolutional neural network (CNN) has made outstanding performance in the classification of natural light images. However, images in many fields have the characteristics of high noise, low resolution, no color information and small data set, such as mammogram, which will affect the accuracy and robustness of the model. In order to improve the classification accuracy and the noise robustness of convolution network for mammogram images, we design a novel classification model based on the new weighted fusion convolution framework. This method has been improved from the following aspects: firstly, we take the place of traditional max-pooling layer with convolution layer with increased step, which achieves the purpose of down-sampling and extracts features more rationally through back-propagation. Secondly, we fuse multi-level feature maps to make full use of the information contained in the shallow levels and deep levels. At the same time, we design a new fusion method to effectively fuse the feature maps from different layers with different sizes. Finally, our model is tested on the mammographic image analysis society (MIAS), which is a mammographic medical image dataset. The experimental results show that the average accuracy of the model is as high as 97.6%, and the convolution layer with increased step has better robustness than the traditional max-pooling layer.

With the rapid development of deep learning, the research and application of micro-expression recognition are more and more extensive. In the existing solutions, the methods using the difference between onset frame and apex frame have high accuracy and low computational cost. But their extraction of dynamic features is not integrated with classification network, resulting in the lack of feedback from classification loss to the extracted dynamic features. In this paper, we propose a novel Displacement Generating Module (DGM) which uses convolution module to generate the displacement feature between onset frame and apex frame instead of traditional optical flow or dynamic image. The new DGM is integrated with the existing LEARNet to form an end-to-end micro-expression recognition network, where the classification loss can feed backward to the parameters of DGM to obtain better displacement features. We also present a random selection method of apex frame to increase the amount of training data, and present a normalization operation for the displacement features with different scales. Our new approach has been tested on SAMM, SMIC, CASME II datasets with LOSO evaluation method, and achieves 0.737 on UF1 and 0.726 on UAR, which is obviously higher than existing networks with optical flow and dynamic imaging techniques.

Due to the particularity of SAR image, existing SAR image classification models often lack strong robustness against noise. Moreover, SAR images are naturally prone to speckle noise and sensitive to observed azimuth. To solve these problems, in this paper, we propose a novel fusion framework in which the convolutional layer with increased stride is used to replace the max pooling layer. Unlike max pooling layer roughly extracts the maximum pixel value in one region as its main feature, which is easy to introduce noise, convolution operation can update the weights and learn features more rationally by back-propagation. It also can achieve the same purpose of down sampling as pooling layers. In order to make full use of feature maps from different layers, our framework fuses the feature vectors extracted from different layers, which helps improve the performance of our classification model. For the problem of overfitting caused by the small MSTAR dataset of SAR images, we replace fully connected layers with convolution layers to relieve the overfitting of the convolution layers by reducing the number of parameters. In order to improve the robustness against observed azimuth angles of the dataset, we adopt the multi-channel calibration and superposition as model's input, which can be used in real flight platform. The extensive experiments conducted on the MSTAR dataset have clearly demonstrated that our framework achieves higher classification accuracy, stronger robustness against noise than other existing methods, as well as its excellent classification performance for the targets of the same category and different subcategories, which is more difficult to be classified.

The road accidents are a common cause to the injury and death of people. As reported by The American National Highway Traffic Safety Administration (NHTSA), the drivers drowsiness accounts for nearly 100,000 accidents per year in the United States. Thus, we present a novel drivers drowsiness detection system in this paper using the techniques of deep learning(DL), mobile computing, wearable device and Electroencephalography (EEG). We employ the deep learning architecture designed by ourselves that can be easily implemented on the mobile phone to detect the drowsiness with a high accuracy. The EEG signal we use is only single channel that can be easily obtained by the wearable device. The EEG signal collector is designed and made by ourselves, which is like a hair band that makes the driver easier and more comfortable to wear it. The whole system mainly consists of two parts: one is the hardware consisting of EEG headband and sensor, the other is software consisting of Android application and web platform. The app contains the fine trained model to make real-time prediction based on the EEG signal and alert the driver, while sending the data to the backend synchronously. The web platform provides an interface for the monitor to observe the condition of the driver. Our system achieved an accuracy of 97.09% detecting the drivers drowsiness, which surpasses the SOTA methods. The model's size and predict latency are also within a smaller scale than present models that make it more applicable to mobile and embedded system.

It is very difficult to directly detect objects on the entire large scale remote sensing image, due to the limited GPU memory. Moreover, there are no objects of interest in most areas of such a huge image, thus a lot of computational costs is wasted in dealing with these vain areas. Therefore, this paper proposes a Cropping Region Proposal Network (CRPN), which includes a weak semantic RPN for quickly locating interesting regions, and a dual-scale strategy for generating effective cropping regions. Cropping regions consist of small and large cropping scales for detecting various-scale objects including very small and very large objects, which is hard for existing methods. CRPN helps to detect effective regions of remote sensing image. Meanwhile, it is also modularized and can be easily connected with mainstream detectors to form an end-to-end detecting framework. Experiments on public DOTA dataset show that our CRPN is effective for filtering invalid regions to greatly reduce the computation burden, and helps to achieve more accurate object detection on large scale remote sensing images.

Segmentation of the left ventricle (LV) is essential for quantitative calculation of clinical indices for analyzing the cardiac contractile function. However, it is challenging to automatically segment small-contour cardiac magnetic resonance (CMR) images for traditional convolutional neural networks (ConvNets) because of their low robustness to scale variation. In this paper, we propose a multi-scale fusion learning method to advance the performance of ConvNets for the LV segmentation. To realize our multi-scale fusion learning, single-scale input and multi-scale output (SIMO) networks are firstly trained to construct a SIMO-based multi-scale fusion network (SIMO-based MSF_Net). The trained SIMO networks produce different-scale coarse prediction results which are then fused into another multi-scale network. Finally, the coarse results are progressively refined to yield finer segmentation results. Our multi-scale fusion learning is evaluated on MICCAI 2009 challenging database for the LV segmentation. Experimental results demonstrate the robustness of our SIMO-based MSF_Net for the segmentation of challenging CMR images and the metric of “Good contours” achieves 98.35% on the testing set, which is greatly improved compared with the state-of-the-art methods.

Haptic-based tissue stiffness perception is essential for palpation training system, which can provide the surgeon haptic cues for improving the diagnostic abilities. However, current haptic devices, such as Geomagic Touch, fail to provide immersive and natural haptic interaction in virtual surgery due to the inherent mechanical friction, inertia, limited workspace and flawed haptic feedback. To tackle this issue, we design a novel magnetic levitation haptic device based on electromagnetic principles to augment the tissue stiffness perception in virtual environment. Users can naturally interact with the virtual tissue by tracking the motion of magnetic stylus using stereoscopic vision so that they can accurately sense the stiffness by the magnetic stylus, which moves in the magnetic field generated by our device. We propose the idea that the effective magnetic field (EMF) is closely related to the coil attitude for the first time. To fully harness the magnetic field and flexibly generate the specific magnetic field for obtaining required haptic perception, we adopt probability clouds to describe the requirement of interactive applications and put forward an algorithm to calculate the best coil attitude. Moreover, we design a control interface circuit and present a self-adaptive fuzzy proportion integration differentiation (PID) algorithm to precisely control the coil current. We evaluate our haptic device via a series of quantitative experiments which show the high consistency of the experimental and simulated magnetic flux density, the high accuracy (0.28 mm) of real-time 3D positioning and tracking of the magnetic stylus, the low power consumption of the adjustable coil configuration, and the tissue stiffness perception accuracy improvement by 2.38 percent with the self-adaptive fuzzy PID algorithm. We conduct a user study with 22 participants, and the results suggest most of the users can clearly and immersively perceive different tissue stiffness and easily detect the tissue abnormality. Experimental results demonstrate that our magnetic levitation haptic device can provide accurate tissue stiffness perception augmentation with natural and immersive haptic interaction.

Uterine fibroids segmentation in ultrasound images is of great importance in the definition of intra‐operative planning of ultrasound‐guided high‐intensity focused ultrasound (HIFU) therapy. However, it is challenging to obtain accurate, robust and efficient uterine fibroid segmentation due to low quality of ultrasound images. In this study, the authors propose a novel adaptive localised region and edge‐based active contour model using shape constraint and sub‐global information to accurately and efficiently segment the uterine fibroids in ultrasound images with robustness against initial contour. The authors first define adaptive local radius for the localised region‐based model and combine it with the edge‐based model to accurately and efficiently capture image's heterogeneous features and edge features. Then, they incorporate a shape constraint to reduce boundary leakage or excessive contraction to obtain more accurate segmentation. To overcome the initialisation sensitivity, they introduce the sub‐global information to prevent the curve from trapping into the local minima and obtain robust results. Furthermore, the authors optimise computation by adaptively sharing local region and employing the multi‐scale segmentation method to achieve efficient segmentation. The proposed method is validated by uterine fibroid ultrasound images in HIFU therapy and the results demonstrate that it can achieve accurate, robust and efficient segmentation.

Abstract Soft tissue warping is one of the key technologies of medical dynamics simulation, such as surgical simulation, image guided surgery. In this paper, we present a novel simulation method which is stable and fast like linear models for soft tissue warping simulation. This method performs on the irregular mesh models, and it is able to represent the visual properties of physical processes with low computational complexity using the Force Asynchronous Diffusion Model (FADM) proposed in this paper. It contains three parts: model preprocessing, collision detection and simulation model solution. In model preprocessing, we establish three models based on the triangular mesh: the geometrical model, the physical model and the transitional model. A two‐level collision detection algorithm is presented based on the three models. At every time step of the simulation model solution, to more accurately reflect the internal physical properties of the soft tissue, we divide the springs in physical model into three kinds: tissue springs, connection springs and virtual springs; and we propose the asynchronous regions and active regions to simplify the computing process according to the realistic physical warping. Experimental results show the FAMD can achieve good warping effects on speed and realism. Copyright © 2011 John Wiley & Sons, Ltd.