Deciphering PDH1's role in mung bean domestication: a genomic perspective on pod dehiscence.

Mung bean (Vigna radiata) stands as a crucial legume crop in Asia, contributing to food security. However, our understanding of the underlying genetic foundation governing domesticated agronomic traits, especially those linked to pod architecture, remains largely unexplored. In this study, we delved into the genomic divergence between wild and domesticated mung bean varieties, leveraging germplasm obtained from diverse sources. Our findings unveiled pronounced variation in promoter regions (35%) between the two mung bean subpopulations, suggesting substantial changes in gene expression patterns during domestication. Leveraging transcriptome analysis using distinct reproductive stage pods and subpopulations, we identified candidate genes responsible for pod and seed architecture development, along with Genome-Wide Association Studies (GWAS) and Quantitative Trait Locus (QTL) analysis. Notably, our research conclusively confirmed PDH1 as a parallel domesticated gene governing pod dehiscence in legumes. This study imparts valuable insights into the genetic underpinnings of domesticated agronomic traits in mung bean, and simultaneously highlighting the parallel domestication of pivotal traits within the realm of legume crops.

ANKRD22 promotes resolution of psoriasiform skin inflammation by antagonizing NIK-mediated IL-23 production.

Inflammation resolution is an essential process for preventing the development of chronic inflammatory diseases. However, the mechanisms that regulate inflammation resolution in psoriasis are not well understood. Here, we report that ANKRD22 is an endogenous negative orchestrator of psoriasiform inflammation because ANKRD22-deficient mice are more susceptible to IMQ-induced psoriasiform inflammation. Mechanistically, ANKRD22 deficiency leads to excessive activation of the TNFRII-NIK-mediated noncanonical NF-κB signaling pathway, resulting in the hyperproduction of IL-23 in DCs. This is due to ANKRD22 being a negative feedback regulator for NIK because it physically binds to and assists in the degradation of accumulated NIK. Clinically, ANKRD22 is negatively associated with IL-23A expression and psoriasis severity. Of greater significance, subcutaneous administration of an AAV carrying ANKRD22-overexpression vector effectively hastens the resolution of psoriasiform skin inflammation. Our findings suggest ANKRD22, an endogenous supervisor of NIK, is responsible for inflammation resolution in psoriasis, and may be explored in the context of psoriasis therapy.

Change of function and brain activity in patients of right spastic arm paralysis combined with aphasia after contralateral cervical seventh nerve transfer surgery.

Left hemisphere injury can cause right spastic arm paralysis and aphasia, and recovery of both motor and language functions shares similar compensatory mechanisms and processes. Contralateral cervical seventh cross transfer (CC7) surgery can provide motor recovery for spastic arm paralysis by triggering interhemispheric plasticity, and self-reports from patients indicate spontaneous improvement in language function but still need to be verified. To explore the improvements in motor and language function after CC7 surgery, we performed this prospective observational cohort study. The Upper Extremity part of Fugl-Meyer scale (UEFM) and Modified Ashworth Scale were used to evaluate motor function, and Aphasia Quotient calculated by Mandarin version of the Western Aphasia Battery (WAB-AQ, larger score indicates better language function) was assessed for language function. In 20 patients included, the average scores of UEFM increased by .40 and 3.70 points from baseline to 1-week and 6-month post-surgery, respectively. The spasticity of the elbow and fingers decreased significantly at 1-week post-surgery, although partially recurred at 6-month follow-up. The average scores of WAB-AQ were increased by 9.14 and 10.69 points at 1-week and 6-month post-surgery (P < .001 for both), respectively. Post-surgical fMRI scans revealed increased activity in the bilateral hemispheres related to language centrals, including the right precentral cortex and right gyrus rectus. These findings suggest that CC7 surgery not only enhances motor function but may also improve the aphasia quotient in patients with right arm paralysis and aphasia due to left hemisphere injuries.

Constructing Lanthanide-Organic Complexes for X-ray Scintillation and Imaging.

The photoluminescent properties of lanthanide complexes have been thoroughly investigated; however, there have been much fewer studies showcasing their potential use in ionizing radiation detection. In this work, we delve into the photo- and radio-induced luminescence of a series of lanthanide-bearing organic-inorganic hybrids and their potential as a platform for X-ray scintillation and imaging. The judicious synergy between lanthanide cations and 2,6-di(1H-pyrazol-1-yl)isonicotinate (bppCOO-) ligands affords six new materials with three distinct structures. Notably, Eu-bppCOO-1 and Tb-bppCOO-2 display sharp fingerprint X-ray-excited luminescence (XEL), the intensities of which can be linearly correlated with the X-ray dose rates over a broad dynamic range (0.007-4.55 mGy s-1). Moreover, the X-ray sensing efficacies of Eu-bppCOO-1 and Tb-bppCOO-2 were evaluated, showing that Tb-bppCOO-2 features a lower detection limit of 4.06 µGy s-1 compared to 14.55 µGy s-1 of Eu-bppCOO-1. Given the higher X-ray sensitivity and excellent radiation stability of Tb-bppCOO-2, we fabricated a flexible scintillator film for X-ray imaging by embedding finely ground Tb-bppCOO-2 in the polydimethylsiloxane (PDMS) polymer. The resulting scintillator film can be utilized for high-resolution X-ray imaging with a spatial resolution of approximately 7 lp mm-1.

Establishment of a risk prediction model for bowel preparation failure prior to colonoscopy.

BACKGROUND: This study aimed to determine the factors that contribute to the failure of bowel preparation in patients undergoing colonoscopy and to develop a risk prediction model. METHODS: A total of 1115 outpatients were included. Patients were randomly divided into two groups: the modeling group (669 patients) and the validation group (446 patients). In the modeling group, patients were further divided into two groups based on their success and failure in bowel preparation using the Boston Bowel Preparation Scale. A logistic regression analysis model was used to determine the risk factors of bowel preparation failure, which was subsequently visualized using an alignment diagram. RESULTS: After controlling for relevant confounders, multifactorial logistic regression results showed that age ≥ 60 years (OR = 2.246), male (OR = 2.449), body mass index ≥ 24 (OR = 2.311), smoking (OR = 2.467), chronic constipation (OR = 5.199), diabetes mellitus (OR = 5.396) and history of colorectal surgery (OR = 5.170) were influencing factors of bowel preparation failure. The area under the ROC curve was 0.732 in the modeling group and 0.713 in the validation group. According to the calibration plot, the predictive effect of the model and the actual results were in good agreement. CONCLUSIONS: Age ≥ 60 years, male, body mass index ≥ 24, smoking, chronic constipation, diabetes mellitus, and history of colorectal surgery are independent risk factors for bowel preparation failure. The established prediction model has a good predictive efficacy and can be used as a simple and effective tool for screening patients at high risk for bowel preparation failure.

Voriconazole plus flucytosine is not superior to amphotericin B deoxycholate plus flucytosine as an induction regimen for cryptococcal meningitis treatment.

BACKGROUND: The efficacy and side effects of voriconazole plus 5-flucytosine (Vori + 5-FC) versus amphotericin B deoxycholate plus 5-flucytosine (AmBd + 5-FC) as an induction treatment for cryptococcal meningitis are unknown. METHODS: Forty-seven patients treated with Vori + 5-FC and 92 patients treated with AmBd + 5-FC were included in the current study after propensity score matching (PSM) at a ratio of 1:2. Two-week laboratory test results and 90-day mortality were compared between the two groups. RESULTS: After 2 weeks of induction treatment, the CSF Cryptococcus sterile culture rate was 57.1% in the Vori + 5-FC group and 76.5% in the AmBd + 5-FC group (p = .026). No difference was found in the normalization of CSF indicators (glucose, total protein, intracranial pressure and India ink sterile rate) between the two groups. Both the Vori + 5FC regimen and AmBd + 5-FC regimen obviously decreased haemoglobin concentrations, platelet counts and serum potassium levels (all p ≤ .010). Notably, the Vori + 5FC regimen did not influence serum creatinine levels (p = .263), while AmBd + 5FC increased serum creatinine levels (p = .019) after 2-week induction treatment. The Vori + 5-FC group and AmBd + 5-FC group had similar 90-day cumulative survival rates (89.9% vs. 87.8%, p = .926). CONCLUSION: The Vori + 5-FC regimen was associated with low 2-week CSF sterile culture and was not superior to AmBd + 5-FC as induction therapy in terms of the 90-day cumulative survival rate of CM patients.

Development and validation of a nomogram for predicting critical respiratory events during early anesthesia recovery in elderly patients.

BACKGROUND: Elderly patients undergoing recovery from general anesthesia face a heightened risk of critical respiratory events (CREs). Despite this, there is a notable absence of effective predictive tools tailored to this specific demographic. This study aims to develop and validate a predictive model (nomogram) to address this gap. CREs pose significant risks to elderly patients during the recovery phase from general anesthesia, making it an important issue in perioperative care. With the increasing aging population and the complexity of surgical procedures, it is crucial to develop effective predictive tools to improve patient outcomes and ensure patient safety during post-anesthesia care unit (PACU) recovery. METHODS: A total of 324 elderly patients who underwent elective general anesthesia in a grade A tertiary hospital from January 2023 to June 2023 were enrolled. Risk factors were identified using least absolute shrinkage and selection operator (LASSO) regression. A multivariate logistic regression model was constructed and represented as a nomogram. Internal validation of the model was performed using Bootstrapping. This study followed the TRIPOD checklist for reporting. RESULTS: The indicators included in the nomogram were frailty, snoring, patient-controlled intravenous analgesia (PCIA), emergency delirium and cough intensity at extubation. The diagnostic performance of the nomogram model was satisfactory, with AUC values of 0.990 and 0.981 for the training set and internal validation set, respectively. The optimal cutoff value was determined to be 0.22, based on a Youden index of 0.911. The F1-score was 0.927, and the MCC was 0.896. The calibration curve, Brier score (0.046), and HL test demonstrated acceptable consistency between the predicted and actual results. DCA revealed high net benefits of the nomogram prediction across all threshold probabilities. CONCLUSIONS: This study developed and validated a nomogram to identify elderly patients in the PACU who are at higher risk of CREs. The identified predictive factors included frailty condition, snoring syndrome, PCIA, emergency delirium, and cough intensity at extubation. By identifying patients at higher risk of CREs early on, medical professionals can implement targeted strategies to mitigate the occurrence of complications and provide better postoperative care for elderly patients recovering from general anesthesia.

Nonlinear Regularization Decoding Method for Speech Recognition.

Existing end-to-end speech recognition methods typically employ hybrid decoders based on CTC and Transformer. However, the issue of error accumulation in these hybrid decoders hinders further improvements in accuracy. Additionally, most existing models are built upon Transformer architecture, which tends to be complex and unfriendly to small datasets. Hence, we propose a Nonlinear Regularization Decoding Method for Speech Recognition. Firstly, we introduce the nonlinear Transformer decoder, breaking away from traditional left-to-right or right-to-left decoding orders and enabling associations between any characters, mitigating the limitations of Transformer architectures on small datasets. Secondly, we propose a novel regularization attention module to optimize the attention score matrix, reducing the impact of early errors on later outputs. Finally, we introduce the tiny model to address the challenge of overly large model parameters. The experimental results indicate that our model demonstrates good performance. Compared to the baseline, our model achieves recognition improvements of 0.12%, 0.54%, 0.51%, and 1.2% on the Aishell1, Primewords, Free ST Chinese Corpus, and Common Voice 16.1 datasets of Uyghur, respectively.

The recent advance of precisely designed membranes for sieving.

Developing new membranes with both high selectivity and permeability is critical in membrane science since conventional membranes are often limited by the trade-off between selectivity and permeability. In recent years, the emergence of advanced materials with accurate structures at atomic or molecular scale, such as metal organic framework, covalent organic framework, graphene, has accelerated the development of membranes, which benefits the precision of membrane structures. In this review, current state-of-the-art membranes are first reviewed and classified into three different types according to the structures of their building blocks, including laminar structured membranes, framework structured membranes and channel structured membranes, followed by the performance and applications for representative separations (liquid separation and gas separation) of these precisely designed membranes. Last, the challenges and opportunities of these advanced membranes are also discussed.

Three-dimensional genome landscape comprehensively reveals patterns of spatial gene regulation in papillary and anaplastic thyroid cancers: a study using representative cell lines for each cancer type.

BACKGROUND: Spatial chromatin structure is intricately linked with somatic aberrations, and somatic mutations of various cancer-related genes, termed co-mutations (CoMuts), occur in certain patterns during cancer initiation and progression. The functional mechanisms underlying these genetic events remain largely unclear in thyroid cancer (TC). With discrepant differentiation, papillary thyroid cancer (PTC) and anaplastic thyroid cancer (ATC) differ greatly in characteristics and prognosis. We aimed to reveal the spatial gene alterations and regulations between the two TC subtypes. METHODS: We systematically investigated and compared the spatial co-mutations between ATC (8305C), PTC (BCPAP and TPC-1), and normal thyroid cells (Nthy-ori-3-1). We constructed a framework integrating whole-genome sequencing (WGS), high-throughput chromosome conformation capture (Hi-C), and transcriptome sequencing, to systematically detect the associations between the somatic co-mutations of cancer-related genes, structural variations (SVs), copy number variations (CNVs), and high-order chromatin conformation. RESULTS: Spatial co-mutation hotspots were enriched around topologically associating domains (TADs) in TC. A common set of 227 boundaries were identified in both ATC and PTC, with significant overlaps between them. The spatial proximities of the co-mutated gene pairs in the two TC types were significantly greater than in the gene-level and overall backgrounds, and ATC cells had higher TAD contact frequency with CoMuts > 10 compared with PTC cells. Compared with normal thyroid cells, in ATC the number of the created novel three-dimensional chromatin structural domains increased by 10%, and the number of shifted TADs decreased by 7%. We found five TAD blocks with CoMut genes/events specific to ATC with certain mutations in genes including MAST-NSUN4, AM129B/TRUB2, COL5A1/PPP1R26, PPP1R26/GPSM1/CCDC183, and PRAC2/DLX4. For the majority of ATC and PTC cells, the HOXA10 and HIF2α signals close to the transcription start sites of CoMut genes within TADs were significantly stronger than those at the background. CNV breakpoints significantly overlapped with TAD boundaries in both TC subtypes. ATCs had more CNV losses overlapping with TAD boundaries, and noncoding SVs involved in intrachromosomal SVs, amplified inversions, and tandem duplication differed between ATC and PTC. TADs with short range were more abundant in ATC than PTC. More switches of A/B compartment types existed in ATC cells compared with PTC. Gene expression was significantly synchronized, and orchestrated by complex epigenetics and regulatory elements. CONCLUSION: Chromatin interactions and gene alterations and regulations are largely heterogeneous in TC. CNVs and complex SVs may function in the TC genome by interplaying with TADs, and are largely different between ATC and PTC. Complexity of TC genomes, which are highly organized by 3D genome-wide interactions mediating mutational and structural variations and gene activation, may have been largely underappreciated. Our comprehensive analysis may provide key evidence and targets for more customized diagnosis and treatment of TC.

Display-Semantic Transformer for Scene Text Recognition.

Linguistic knowledge helps a lot in scene text recognition by providing semantic information to refine the character sequence. The visual model only focuses on the visual texture of characters without actively learning linguistic information, which leads to poor model recognition rates in some noisy (distorted and blurry, etc.) images. In order to address the aforementioned issues, this study builds upon the most recent findings of the Vision Transformer, and our approach (called Display-Semantic Transformer, or DST for short) constructs a masked language model and a semantic visual interaction module. The model can mine deep semantic information from images to assist scene text recognition and improve the robustness of the model. The semantic visual interaction module can better realize the interaction between semantic information and visual features. In this way, the visual features can be enhanced by the semantic information so that the model can achieve a better recognition effect. The experimental results show that our model improves the average recognition accuracy on six benchmark test sets by nearly 2% compared to the baseline. Our model retains the benefits of having a small number of parameters and allows for fast inference speed. Additionally, it attains a more optimal balance between accuracy and speed.

Scene Uyghur Recognition Based on Visual Prediction Enhancement.

Aiming at the problems of Uyghur oblique deformation, character adhesion and character similarity in scene images, this paper proposes a scene Uyghur recognition model with enhanced visual prediction. First, the content-aware correction network TPS++ is used to perform feature-level correction for skewed text. Then, ABINet is used as the basic recognition network, and the U-Net structure in the vision model is improved to aggregate horizontal features, suppress multiple activation phenomena, better describe the spatial characteristics of character positions, and alleviate the problem of character adhesion. Finally, a visual masking semantic awareness (VMSA) module is added to guide the vision model to consider the language information in the visual space by masking the corresponding visual features on the attention map to obtain more accurate visual prediction. This module can not only alleviate the correction load of the language model, but also distinguish similar characters using the language information. The effectiveness of the improved method is verified by ablation experiments, and the model is compared with common scene text recognition methods and scene Uyghur recognition methods on the self-built scene Uyghur dataset.

Lightweight Scene Text Recognition Based on Transformer.

Scene text recognition (STR) has been a hot research field in computer vision, aiming to recognize text in natural scenes using computers. Currently, attention-based encoder-decoder frameworks struggle to precisely align feature regions with the target object when dealing with complex and low-quality images, a phenomenon known as attention drift. Additionally, with the rise of Transformer, the increasing size of parameters results in higher computational costs. In order to solve the above problems, based on the latest research results of Vision Transformer (ViT), we utilize an additional position-enhancement branch to alleviate attention drift and dynamically fused position information with visual information to achieve better recognition accuracy. The experimental results demonstrate that our model achieves a 3% higher average recognition accuracy on the test set compared to the baseline. Meanwhile, our model maintains the advantage of a small number of parameters and fast inference speed, achieving a good balance between accuracy, speed, and computational load.

Preclinical and clinical study on [18F]DRKXH1: a novel ß-amyloid PET tracer for Alzheimer's disease.

BACKGROUND: The deposition of ß-amyloid (Aß) in the brain is a biomarker of Alzheimer's disease (AD). Highly sensitive Aß positron emission tomography (PET) imaging plays an essential role in diagnosing and evaluating the therapeutic effects of AD. AIM: To synthesize a new Aß tracer [18F]DRKXH1 (5-(4-(6-(2-[18]fluoroethoxy)ethoxy)imidazo[1,2-alpha]pyridin-2-yl)phenyl) and evaluate the tracer performance by biodistribution analysis, in vivo small-animal PET-CT dynamic scan, ex vivo and in vitro autoradiography, and PET in human subjects. METHODS: [18F]DRKXH1 was synthesized automatically by the GE FN module. Log D (pH 7.4) and biodistribution of [18F]DRKXH1 were investigated. Small-animal-PET was used for [18F]DRKXH1 and [18F]AV45 imaging study in AD transgenic mice (APPswe/PSEN1dE9) and age-matched normal mice. The distribution volume ratios (DVR) and standardized uptake value ratios (SUVRs) were calculated with the cerebellum as the reference region. The deposition of Aß plaques in the brain of AD transgenic mice was determined by ex vivo autoradiography and immunohistochemistry. In vitro autoradiography was performed in the postmortem brain sections of AD patients and healthy controls. Two healthy control subjects and one AD patient was subjected to in vivo PET study using [18F]DRKXH1. RESULTS: The yield of [18F]DRKXH1 was 40%, and the specific activity was 156.64 ± 11.55 GBq/µmol. [18F]DRKXH1 was mainly excreted through the liver and kidney. The small-animal PET study showed high initial brain uptake and rapid washout of [18F]DRKXH1. The concentration of [18F]DRKXH1 was detected in the cortex and hippocampus of AD transgenic mice brain. The cortex DVR of AD transgenic mice was higher than that of WT mice (P < 0.0001). Moreover, the SUVRs of AD transgenic mice were higher than those of WT mice based on the 0-60-min dynamic scanning. In vitro autoradiography showed a significant concentration of tracer in the Aß plaque-rich areas in the brain of AD transgenic mice. The DVR value of [18F]-DRKXH1 is higher than that of [18F]-AV45 (1.29 ± 0.05 vs. 1.05 ± 0.08; t = 5.33, P = 0.0003). Autoradiography of postmortem human brain sections showed [18F]DRKXH1-labeled Aß plaques in the AD brain. The AD patients had high retention in cortical regions, while healthy control subjects had uniformly low radioactivity uptake. CONCLUSIONS: [18F]DRKXH1 is an Aß tracer with high sensitivity in preclinical study and has the potential for in vivo detection of the human brain.

Distal radial artery as an alternative approach to forearm radial artery for perioperative blood pressure monitoring: a randomized, controlled, noninferiority trial.

BACKGROUND: The novel distal radial artery (dRA) approach is a popular arterial access route for interventional cardiology and neurointerventions. We explored the dRA as an alternative site to the classic forearm radial artery (RA) for perioperative blood pressure monitoring. We hypothesized that dRA catheterization is noninferior to RA for the first attempt success rate. METHODS: This was a single-center, prospective, randomized controlled, noninferiority study. Adult patients who underwent elective surgery at the Jinling Hospital from May 2021 to August 2021 were enrolled. The primary endpoint was to test the noninferiority of the first attempt success rate between the groups. Secondary endpoints included anatomical characteristics, catheterization time, arterial posterior wall puncture rate, postoperative compression time, dampened arterial pressure waveforms, and complications. RESULTS: Totally, 161 patients who received either dRA (n = 81) or RA (n = 80) catheterization were analyzed. The first attempt success rates were 87.7 and 91.3% in the dRA and RA groups, respectively, with a mean difference of - 3.6% (95% CI, - 13.1 to 5.9%). The dRA diameter and cross-sectional area were significantly smaller than those of the RA (P < 0.001). The subcutaneous depth of dRA was significantly greater than that of the RA (P < 0.001). The dRA had a longer catheterization time (P = 0.008) but a shorter postoperative compression time (P < 0.001). The arterial posterior wall puncture rate of dRA was significantly higher than that of the RA (P = 0.006). The dRA had fewer dampened arterial waveforms than RA (P = 0.030) perioperatively. CONCLUSIONS: The dRA is a rational alternative approach to RA for perioperative arterial pressure monitoring and provides a noninferior first attempt success rate. TRIAL REGISTRATION: This study is registered in the Chinese Clinical Trials Registry (registration number: ChiCTR2100043714 , registration date: 27/02/2021).

Pupil maximum constriction velocity predicts post-induction hypotension in patients with lower ASA status: a prospective observational study.

BACKGROUND: Individuals affected by autonomic dysfunction are at a higher risk of developing hypotension following anesthesia induction. Dynamic pupillometry has previously been employed as a means of assessing autonomic function. This prospective observational study was developed to determine whether pupillary light reflex (PLR) parameters can reliably predict post-induction hypotension (PIH). METHODS: This study enrolled patients with lower ASA status (I-II) undergoing elective surgery. PLR recordings for these patients prior to anesthesia induction were made with an infrared pupil camcorder, with a computer being used to assess Average Constriction Velocity (ACV), Maximum Constriction Velocity (MCV), and Constriction Ratio (CR). PIH was defined by a > 30% reduction in mean arterial pressure (MAP) or any MAP recording < 65 mmHg for at least 1 min from the time of induction until 10 minutes following intubation. Patients were stratified into PIH and non-PIH groups based on whether or not they developed hypotension. RESULTS: This study enrolled 61 total patients, of whom 31 (50.8%) exhibited one or more hypotensive episodes. Patients in the PIH group exhibited significantly smaller ACV (P = 0.003) and MCV values (P < 0.001), as well as a higher CR (P = 0.003). Following adjustment for certain factors (Model 2), MCV was identified as a protective factor for PIH (Odds Ratio: 0.369). Receiver operating characteristic (ROC) analyses revealed that relative to CR (AUC: 0.695, 95% CI: 0.563-0.806; P = 0.004), the reciprocal of MCV (1/MCV) offered greater value as a predictor of PIH (AUC: 0.803,95%CI: 0.681-0.894; P < 0.001). CONCLUSION: These results indicate that pupil maximum constriction velocity is a reliable predictor of post-induction hypotension in individuals of ASA I-II status undergoing elective surgery. TRIAL REGISTRATION: This study was registered with the Chinese Clinical Trial Registry (registration number: ChiCTR2200057164, registration date: 01/03/2022).

A Robust Method: Arbitrary Shape Text Detection Combining Semantic and Position Information.

There is a growing interest in scene text detection for arbitrary shapes. The effectiveness of text detection has also evolved from horizontal text detection to the ability to perform text detection in multiple directions and arbitrary shapes. However, scene text detection is still a challenging task due to significant differences in size and aspect ratio and diversity in shape, as well as orientation, coarse annotations, and other factors. Regression-based methods are inspired by object detection and have limitations in fitting the edges of arbitrarily shaped text due to the characteristics of their methods. Segmentation-based methods, on the other hand, perform prediction at the pixel level and thus can fit arbitrarily shaped text better. However, the inaccuracy of image text annotations and the distribution characteristics of text pixels, which contain a large number of background pixels and misclassified pixels, degrades the performance of segmentation-based text detection methods to some extent. Usually, considering whether a pixel belongs to a text region is highly dependent on the strength of the semantic information it has and the position of the pixel in the text area. Based on the above two points, we propose an innovative and robust method for scene text detection combining position and semantic information. First, we add position information to the images using a position encoding module (PosEM) to help the model learn the implicit feature relationships associated with the position. Second, we use the semantic enhancement module (SEM) to enhance the model's focus on the semantic information in the image during feature extraction. Then, to minimize the effect of noise due to inaccurate image text annotations and the distribution characteristics of text pixels, we convert the detection results into a probability map that can more reasonably represent the text distribution. Finally, we reconstruct and filter the text instances using a post-processing algorithm to reduce false positives. The experimental results show that our model improves significantly on the Total-Text, MSRA-TD500, and CTW1500 datasets, outperforming most previous advanced algorithms.

Attention-Based Scene Text Detection on Dual Feature Fusion.

The segmentation-based scene text detection algorithm has advantages in scene text detection scenarios with arbitrary shape and extreme aspect ratio, depending on its pixel-level description and fine post-processing. However, the insufficient use of semantic and spatial information in the network limits the classification and positioning capabilities of the network. Existing scene text detection methods have the problem of losing important feature information in the process of extracting features from each network layer. To solve this problem, the Attention-based Dual Feature Fusion Model (ADFM) is proposed. The Bi-directional Feature Fusion Pyramid Module (BFM) first adds stronger semantic information to the higher-resolution feature maps through a top-down process and then reduces the aliasing effects generated by the previous process through a bottom-up process to enhance the representation of multi-scale text semantic information. Meanwhile, a position-sensitive Spatial Attention Module (SAM) is introduced in the intermediate process of two-stage feature fusion. It focuses on the one feature map with the highest resolution and strongest semantic features generated in the top-down process and weighs the spatial position weight by the relevance of text features, thus improving the sensitivity of the text detection network to text regions. The effectiveness of each module of ADFM was verified by ablation experiments and the model was compared with recent scene text detection methods on several publicly available datasets.

Emergence of Thorium-Based Polyoxo Clusters as a Platform for Selective X-ray Dosimetry.

A brand-new application of thorium-bearing clusters in the field of ionizing radiation detection is exemplified by two novel hexanuclear thorium clusters, Th-bppCOO-1 and Th-bppCOO-2, which incorporate carboxylate-functionalized 2,6-di(pyrazol-1-yl)pyridine ligands. Notably, Th-bppCOO-1 is composed of an unprecedented [Th6(OH)4O4(H2O)5]12+ secondary building unit, the Th6 core of which is decorated by five H2O molecules. Furthermore, selective photoluminescence quenching responses of Th-bppCOO-1 and Th-bppCOO-2 toward X-ray over UV radiation have been demonstrated for the first time.

Geometrical-lightguide-based head-mounted lightfield displays using polymer-dispersed liquid-crystal films.

Integrating the promising waveguide or lightguide optical combiners to head-mounted lightfield display (LF-HMD) systems offers a great opportunity to achieve both a compact optical see-through capability required for augmented or mixed reality applications and true 3D scene with correct focus cues required for mitigating the well-known vergence-accommodation conflict. Due to the non-sequential ray propagation nature of these flat combiners and the ray construction nature of a lightfield display engine, however, adapting these two technologies to each other confronts several significant challenges. In this paper, we explore the feasibility of combining an integral-imaging-based lightfield display engine with a geometrical lightguide based on microstructure mirror arrays. The image artifacts and the key challenges in a lightguide-based LF-HMD system are systematically analyzed and are further quantified via a non-sequential ray tracing simulation. We further propose to utilize polymer-dispersed liquid-crystal (PDLC) films to address the inherent problems associated with a lightguide combiner such as increasing the viewing density and improving the image coupling uniformity. We finally demonstrate, to our best knowledge, the first lightguide-based LF-HMD system that takes the advantages of both the compact form factor of a lightguide combiner and the true 3D virtual image rendering capability of a lightfield display.