Model generalization to the unseen scenes is crucial real-world applications, such as autonomous driving, which requires robust vision systems. To enhance model generalization, domain through learning domain-invariant representation has been widely studied. However, most existing works learn shared feature space within multi-source domains but ignore characteristic of itself (e.g., sensitivity domain-specific style). Therefore, we propose Domain-invariant Representation Learning (DIRL) for...

Ultrasound examination is widely used for diagnosis of breast cancer, which requires a full-coverage scan the whole organ and stable probe–breast interaction high-quality image acquisition. In this study, robotic system automated ultrasound scanning proposed. To avoid occlusion, point cloud obtained from multiple angles registered together accurate tissue shape reconstruction, then path planning performed with isometric 3-D searching algorithm full uniform coverage. reduce adverse influence...

Robotic breast ultrasound (RBUS) aims to standardize ultrasonography, reduce the workload of sonographers, and provide high-quality (US) images for subsequent diagnosis. In process RBUS screening, adjusting US probe correctly efficiently acquire is fundamental significant. this letter, a learning-based adjustment framework proposed. Firstly, lightweight multi-task combination approach utilized jointly assessing imaging quality with multiple indicators. Then, an experience-guided learning...

Autonomous ultrasound scanning robots have attracted the attention of researchers, and real-time quality assessment images is key technology them. Existing robot systems usually use pixel-level feature statistical methods such as grayscale, confidence map, etc. However, in clinical practice doctors' evaluation image not only relies on pixel quality, but also content. In this study, we introduced deep learning method to medical breast learn standards. We collected 1205 533 patients asked...

Offline reinforcement learning (ORL) has been getting increasing attention in robot learning, benefiting from its ability to avoid hazardous exploration and learn policies directly precollected samples. Approximate policy iteration (API) is one of the most commonly investigated ORL approaches robotics, due linear representation policies, which makes it fairly transparent both theoretical engineering analysis. One open problem API how design efficient effective basis functions. The broad...

Since the development of optic-fiber interferometers and design outstanding speech recognition models, study on perimeter intrusion detection systems (PIDS) becomes a field interest. In this paper, an based fence system that uses Sagnac mixture Gaussian hidden Markov models (GMM-HMMs) is proposed. Experiments real intrusions are performed, comparisons also carried out between our approach SVM method, which prove more robust accurate.

Purpose: To isolate and identify the chemical components of Paraoncidium reevesii.Methods: Silica gel column chromatography was used to from petroleum ether ethyl acetate fractions acetone extract, structures compounds were derived 1H-nuclear magnetic resonance (1H-NMR), 13C-nuclear (13C-NMR) mass spectrometry (MS) analyses also with aid literature data for authenticated samples.Results: Cholesterol (1), baconipyrone D (2), chimyl alcohol (3), batyl (4), á-monpalmitin (5), stearic acid (6),...

Traffic safety has become a primary focus on the rapid development of railway traffic. Failure turnout shall endanger train operations and affect their efficiency. Currently, fault diagnosis still relies experience maintenance personnel that can introduce several problems, such as low efficiency large amounts required labor. To solve these this paper takes most widely used ZD6 system in China research object, failure modes are summarized into eight typical types. Combined with actual data,...

Abstract To fulfill intelligent human-machine interface, the crucial information about human health condition tends to be acquired with great sensitivity, stability and durability. Despite some remarkable progress for bioelectric signal perception in recent years, epidermal electrode still suffers from motion artifacts, inability sweat permeability. So evolution of high-performance has been research focus receives much attention years. satisfy growing pursuit smarter sensing, this article...

Turnout is an important part of the railway signal system and also equipment to ensure safety traffic. In view current situation low efficiency using computer monitoring locate faults, this paper combines theory fuzzy cognitive map with fault diagnosis turnout, uses real-coded genetic algorithm which tells initial weight network construct a classifier model for classify faults turnouts. The simulation experiments show that based on can effectively turnout. Compared commonly used classifiers...

Recently, there is growing attention on one-stage panoptic segmentation methods which aim to segment instances and stuff jointly within a fully convolutional pipeline efficiently. However, most of the existing works directly feed backbone features various heads ignoring demands for semantic instance are different: The former needs semantic-level discriminative features, while latter requires be distinguishable across instances. To alleviate this, we propose first predict instance-level...

Acute kidney injury (AKI) is a common clinical syndrome characterized by sudden episode of failure or damage within few hours days. Accurate early prediction AKI for patients in ICU who are more likely than others to have can enable timely interventions, and reduce the complications AKI. Much information relevant captured notes that largely unstructured text requires advanced natural language processing (NLP) useful extraction. On other hand, pre-trained contextual models such as...

Even micromachines with tailored functionalities enable targeted therapeutic applications in biological environments, their controlled motion media and drug delivery functions usually require sophisticated designs complex propulsion apparatuses for practical applications. Covalent organic frameworks (COFs), new chemically versatile nanoporous materials, offer microscale multi-purpose solutions, which are not explored light-driven micromachines. We describe compare two different types of...

Friend recall is an important way to improve Daily Active Users (DAU) in online games. The problem generate a proper lost friend ranking list essentially. Traditional methods focus on rules like intimacy or training classifier for predicting players' return probability, but ignore feature information of (active) players and historical events. In this work, we treat as link prediction explore several which can use features both active players, well Furthermore, propose novel Edge Transformer...

One of the challenges in ultrasound image-guided liver puncture intervention is patient's physiological movement. The respiratory motion modeling basis for accurate positioning target and precise control. existing method estimating among images speckle decorrelation. However, traditional methods require complicated operations are time-consuming, which cannot meet real-time requirement clinical use. In this study, we proposed a learning based with deep neural network to calculate out-of-plane...