Rain streaks in the air appear various blurring degrees and resolutions due to different distances from their positions camera. Similar rain patterns are visible a image as well its multi-scale (or multi-resolution) versions, which makes it possible exploit such complementary information for streak representation. In this work, we explore collaborative representation perspective of input scales hierarchical deep features unified framework, termed progressive fusion network (MSPFN) single...

Video satellite imagery is a new technique for earth dynamic observation and has wide range of uses in environmental fields. Despite its capability targets' detection, it sustains serious restriction the image quality due to degradation compression imaging process. Hence, super-resolution (SR) reconstruction on these compressed low-spatial-resolution images significance afterward ground objects recognition detection tasks. Based recent proposed state-of-the-art convolutional neural networks...

Convolutional neural networks (CNNs) have wide applications in pattern recognition and image processing. Despite recent advances, much remains to be done for CNNs learn a better representation of samples. Therefore, constant optimizations should provided on CNNs. To achieve good performance classification, intuitively, samples' interclass separability, or intraclass compactness simultaneously maximized. Accordingly, this paper, we propose new network, named separability network (SCNet)...

Diffusion probabilistic models (DPMs) which employ explicit likelihood characterization and a gradual sampling process to synthesize data, have gained increasing research interest. Despite their huge computational burdens due the large number of steps involved during sampling, DPMs are widely appreciated in various medical imaging tasks for high-quality diversity generation. Magnetic resonance (MRI) is an important modality with excellent soft tissue contrast superb spatial resolution,...

In a clinical setting, the acquisition of certain medical image modality is often unavailable due to various considerations such as cost, radiation, etc. Therefore, unpaired cross-modality translation techniques, which involve training on data and synthesizing target with guidance acquired source modality, are great interest. Previous methods for images establish one-shot mapping through generative adversarial networks (GANs). As promising alternatives GANs, diffusion models have recently...

Rain streaks in the air appear various blurring degrees and resolutions due to different distances from their positions camera. Similar rain patterns are visible a image as well its multi-scale (or multi-resolution) versions, which makes it possible exploit such complementary information for streak representation. In this work, we explore collaborative representation perspective of input scales hierarchical deep features unified framework, termed progressive fusion network (MSPFN) single...

Hepatic stellate cell (HSC) line LX-2 is activated by liver cancer stem-like cells (LCSLCs) and produces various cytokines that make up most of the hepatocellular carcinoma (HCC) microenvironment. The new genistein derivative, 7-difluoromethoxyl-5,4'-di-n-octylgenistein (DFOG), shows anticancer effects in multiple malignancies controlling forkhead box M1 (FOXM1). In this study, we aimed to assess whether DFOG disrupts crosstalk between human HSC LCSLCs. Distinct generations MHCC97H-derived...

Airborne light detection and ranging (LiDAR) remote sensing enables accurate estimation monitoring of terrain vegetation, digital surface model (DSM) elevation (DEM) are vital analytical tools to achieve this monitoring. Among them, DSM can be directly acquired from airborne LiDAR point clouds; nevertheless, for the production DEM, clouds representing a ground objects should accurately filtered out at first. In some mountain forest areas, due limited penetration LiDAR, points sustain serious...

Previous work for subpixel level Digital Surface Model (DSM) generation mainly focused on data fusion techniques, which are extremely limited by the difficulty of multisource acquisition. Although several DSM super resolution (SR) methods have been developed to ease problem, a new issue that plenty samples needed train model is raised. Therefore, considering original images vital influence its DSM's accuracy, we address problem directly improving resolution. Several SR models refined and...

Both mesenchymal stromal cells (MSCs) and myeloid-derived suppressor (MDSCs) have immunosuppressive properties, their presence may confer a worse prognosis upon cancer patients. However, whether MSCs can enhance the effects of MDSCs in MM remains unknown. We evaluated influence on growth, apoptosis, functions. Our results show that promote proliferation inhibit apoptosis MDSCs. Additionally, ability by inhibiting T-cell IFN-γ production. Furthermore, both mRNA protein levels Arg1 NOS2 were...

In the field of healthcare, acquisition sample is usually restricted by multiple considerations, including cost, labor- intensive annotation, privacy concerns, and radiation hazards, therefore, synthesizing images-of-interest an important tool to data augmentation. Diffusion models have recently attained state-of-the-art results in various synthesis tasks, embedding energy functions has been proved that can effectively guide pre-trained model synthesize target samples. However, we notice...

Signal transducer and activator of transcription 3 (STAT3) is a member the family latent cytoplasmic transcriptional factors that could regulate cell proliferation, survival, development. It has been reported Twist target gene STAT3, STAT3/Twist signaling plays an important role in regulating cancer progress. Here, to explore whether 8-bromo-7-methoxychrysin (BrMC) inhibits liver stem-like (LCSLC) properties via disrupting signaling, we cultured SMMC-7721 cells vitro, evaluated effects BrMC...

Reducing radiation dose in cardiac catheter-based X-ray procedures increases safety but also image noise and artifacts. Excessive artifacts can compromise vital information, which affect clinical decision-making. Developing more effective denoising methodologies will be beneficial to both patients healthcare professionals by allowing imaging at lower without compromising information. This paper proposes a framework based on convolutional neural network (CNN), namely Ultra-Dense Denoising...

In this paper, we propose the novel text mining algorithm based on deep neural network. knowledge representation, capable of embodying information content characteristic and external not only has semantic meaning it is interlinked, characteristics constitute association reveals base. This article through to contents a appearance general different combination co-occurrence analysis, explore analysis method spatial distribution, time distribution application internal mapping mining. terms...

Coal and rock recognition (CRR) has important theoretical practical significance in unmanned coal mining. Laser-induced breakdown spectroscopy (LIBS) is considered a cutting-edge technology the field of material analysis due to its real-time capability, minimal no sample preparation scheme, high sensitivity low atomic weight elements, ability perform nearby distant detection. In this research, new fast accurate coal-rock method for mining based on LIBS presented. The situ sampling face...

Reducing X-ray dose increases safety in cardiac electrophysiology procedures but also image noise and artifacts which may affect the discernibility of devices anatomical cues. Previous denoising methods based on convolutional neural networks (CNNs) have shown improvements quality low-dose fluoroscopy images compromise clinically important details required by cardiologists.

Convolutional neural networks (CNNs) have shown great advantages in computer vision fields, and loss functions are of significance to their gradient descent algorithms. Softmax loss, a combination cross-entropy function, is the most commonly used one for CNNs. Hence, it can continuously increase discernibility sample features classification tasks. Intuitively, promote discrimination CNNs, learned desirable when inter-class separability intra-class compactness maximized simultaneously. Since...

Segmentation is an important prerequisite for developing model healthcare systems, particularly disease diagnosis and treatment planning. In the field of medical image segmentation, U-shaped architecture, commonly referred to as U-Net, has emerged de facto standard achieved remarkable success. However, due intrinsic locality convolution operations, U-Net generally demonstrates limitations in explicitly modeling long-range dependency. Recent transformer-based models, designed...

Digital Elevation Model (DEM), representing the height of earth terrain, is one crucial geographic information products. One main data source DEM airborne LiDAR point cloud with its non-ground-reflections filtered out. Point filtering in thick-forested areas difficult without enough ground control points when using conventional methods. In this paper, a supervised method proposed to handle problem automatic extraction little points. The design inspired by successful application convolutional...

The utilization of optical coherence tomography (OCT) holds significant promise in the realm cervical cancer screening. This study considers characteristics OCT, such as sample scarcity and noise, introduces a noise adaptive diffusion model (NADM) for data augmentation. NADM comprises two components: Adaptive Blind De-noising Module (ABD) Class-Guided Diffusion Model (CDM). CDM is responsible generating high-quality samples, while ABD designed to adaptively suppress noise. Extensive...