Search

Document localization is a promising step for document-based optical character recognition. This task gains difficulty when documents are located in complex natural scene images. In this paper, we propose a coarse-to-fine document localization approach to detect the four corner points of the document in natural scene images. In the first stage, the four corners are roughly predicted through a deep neural networks-based Joint Corner Detector (JCD) with an attention mechanism, which roughly localize the document region via an attentional map. As a key to produce accurate inference of corners, the JCD module suppresses the interference from background in convolutional features substantially. In the second stage, a corner-specific refiner module is designed to refine the previously predicted corners. Considering the different characteristics of the four document corners, the patches cropped around the predicted corners are input into four different corner-specified CNN models, to search the accurate corner locations recursively. Three datasets (ICDAR 2015 SmartDoc competition 1 dataset, SEECS-NUSF dataset and a self-collected dataset) are used to evaluate the performance of our method. The experimental results demonstrate the superiority of the proposed method in localizing the document in natural images, especially in those with complex background. Compared with the state-of-the-art works, our method outperforms most of them.

In 2020, Australia's successful COVID‐19 public health restrictions comprised a national “initial lockdown” (March–May) and “ongoing lockdown” (July–November) for metropolitan Victorian residents only. We evaluated associations between ongoing lockdown and family finances and mental health. In the June and September 2020 Royal Children's Hospital National Child Health Polls, caregivers of children in Victoria and New South Wales (NSW) reported the following: job/income loss; material deprivation (inability to pay for essential items); income poverty; mental health (Kessler‐6); perceived impact on caregiver/child mental health; and caregiver/child coping. Data from caregivers (N= 1207/902) in June/September were analysed using difference‐in‐difference modelling (NSW provided the comparator). During Victoria's ongoing lockdown, job/income loss increased by 11% (95%CI: 3%–18%); Kessler‐6 poor mental health by 6% (95%CI: −0.3%–12%) and perceived negative mental health impacts by 14% for caregivers (95%CI: 6%–23%) and 12% for children (95%CI: 4%–20%). Female (vs. male) caregivers, metropolitan (vs. regional/rural) families, and families with elementary school‐aged children (vs. pre‐/high‐school) were the most affected. The ongoing lockdown was associated with negative experiences of mental health, employment and income, but not deprivation or poverty, likely because of government income supplements introduced early in the pandemic. Future lockdowns require planned responses to outbreaks and evidence‐informed financial and mental health supports.

High-quality interpretation of BRCA1/2variants plays a critical role in the clinical practice of precision medicine. However, a comprehensive system to evaluate the quality and accuracy of variant interpretation has yet to be established. This study investigates the performance of an interpretation system in evaluating the capacities of BRCA1/2interpretation among distinct laboratories in China. The evaluation system is based on a reference database that contains 750 different variants in BRCA1/2. Evaluation was performed among 41 laboratories in China. We classified their performance into five levels. Only level A was considered qualified. This level allows for a 0.3% error rate for clinical decision-related misinterpretation; 26 of 41 laboratories (63%) met the qualified standard, while 7 laboratories were at levels D and E, which indicated egregious mistakes and systemic problems in variant interpretation. Due to strict quality demands, the interpretation of several variants was amended, which largely influenced the quality rate. The number of qualified laboratories would decrease from 26 to 17 if those incorrect recommended interpretations were not corrected. This evaluation system provides a potential approach for standardisation of variant interpretation and lowers the discordance of variant interpretation between different laboratories. A well-designed interpretation ability evaluation is essential to evaluate the interpretation level of laboratories before they provide service in real-world clinical settings.

In this work, we study character counting in STR from a new viewpoint, giving a principled framework showing that the counting information is involved in both visual decoding and semantic decoding. Based on the principled framework, we propose a novel scene text recognizer with a dual character counting-aware visual and semantic modeling network, where the counting information is fused in both vision and language branches. Experimental results demonstrate the effectiveness of our model.

A novel bifunctional MOF-encapsulated cobalt-doped carbon dots nanozyme (Co-CD/PMOF) with excellent peroxidase-mimic catalytic activity and fluorescence property was synthesized and employed to fabricate a chemiluminescence/fluorescence (CL/FL) dual-mode immunosensor for AFB1 detection. Co-CD/PMOF could catalyze the luminol/H2O2system to generate robust and long-lasting CL signals due to the slow diffusion effect and continuous generation of •OH, O2•–, and 1O2species. Differing from traditional flash-type CL emissions, this glow-type CL emission is helpful to fabricate a sensitive and accurate CL sensing platform. Then the CL/FL dual-mode detection of AFB1 was developed using antibody-functionalized Co-CD/PMOF as the signal-amplifying nanoprobe. The CL mode assay based on indirect competitive immune principle was carried out on a chemiluminescence optical fiber platform, where the AFB1-OVA-functionalized optical fiber probe was employed for biorecognition, separation, and signal conducting. The AFB1 detection range and LOD were 0.63–69.36 ng/mL and 0.217 ng/mL, respectively. Using AFB1 antibody-functionalized immunomagnetic beads for capturing and separation, the FL mode detection of AFB1 was established based on the sandwich immune principle. A linear range of 0.54–51.91 ng/mL and a LOD of 0.027 ng/mL were obtained. This work designed a sensitive, rapid, and reliable nanozyme-powered dual-mode assay strategy and provided technical support in the field of environmental monitoring and food safety.

Convolutional neural networks (CNNs) have had great success with regard to the object classification problem. For character classification, we found that training and testing using accurately segmented character regions with CNNs resulted in higher accuracy than when roughly segmented regions were used. Therefore, we expect to extract complete character regions from scene images. Text in natural scene images has an obvious contrast with its attachments. Many methods attempt to extract characters through different segmentation techniques. However, for blurred, occluded, and complex background cases, those methods may result in adjoined or over segmented characters. In this paper, we propose a scene word recognition model that integrates words from small pieces to entire after-cluster-based segmentation. The segmented connected components are classified as four types: background, individual character proposals, adjoined characters, and stroke proposals. Individual character proposals are directly inputted to a CNN that is trained using accurately segmented character images. The sliding window strategy is applied to adjoined character regions. Stroke proposals are considered as fragments of entire characters whose locations are estimated by a stroke spatial distribution system. Then, the estimated characters from adjoined characters and stroke proposals are classified by a CNN that is trained on roughly segmented character images. Finally, a lexicon-driven integration method is performed to obtain the final word recognition results. Compared to other word recognition methods, our method achieves a comparable performance on Street View Text and the ICDAR 2003 and ICDAR 2013 benchmark databases. Moreover, our method can deal with recognizing text images of occlusion and improperly segmented text images.

Highly sensitive and selective immunosensors are widely applied to detect small molecules. Organic solvent extraction of small molecules in samples is generally required when the small molecules have poor solubility in water. Immunosensing assays should tolerate at least some organic solvents to rapidly detect small molecules on-site. Herein, the effects of four organic solvents (methanol, acetone, acetonitrile, and n-hexane) on the homogenous and heterogeneous interactions between two small molecules (bisphonel A (BPA) and aflatoxin M1(AFM1)) and their respective antibodies were studied using an optofluidic immunosensing platform. To evaluate the influences of these solvents on immunoassays for small molecules, indirect competitive immunoassays of BPA and AFM1were conducted using organic solvents at various concentrations. Antibody–antigen binding reaction and inhibition depends on the concentration and water-miscibility of organic solvents as well as their chemical features. The homogenous and heterogeneous binding reactions of small molecules and their antibodies were almost uninfluenced by water-immiscible n-hexane even at a concentration of 40% (v/v). The quantitative immunoassays of small molecules could be carried out in water-miscible organic solvents at low concentrations, including methanol, acetone, and acetonitrile. However, increasing the concentrations of these organic solvents variably affected the immunoassays of the small molecules. The effects of the three water-miscible organic solvents at relatively high concentrations on homogenous binding reactions are more significant than the effects on heterogeneous binding reactions. The homogenous binding reactions between small molecules and their antibodies were inhibited completely by acetone and acetonitrile solutions at high concentrations. The susceptibilities of the binding reactions between different small molecules and their antibodies in response to different organic solvents were variable. To rapidly detect small molecules on-site, an organic solvent should be carefully selected on the basis of its extraction efficiency and effect on the immunoassay.

Tissue homeostasis requires somatic stem cell maintenance; however, mechanisms regulating this process during organogenesis are not well understood. Here, we identify asymmetrically renewing basal and luminal stem cells in the mammary end bud. We demonstrate that SLIT2/ROBO1 signaling regulates the choice between self-renewing asymmetric cell divisions (ACDs) and expansive symmetric cell divisions (SCDs) by governing Inscuteable(mInsc), a key member of the spindle orientation machinery, through the transcription factor Snail (SNAI1). Loss of SLIT2/ROBO1 signaling increases SNAI1 in the nucleus. Overexpression of SNAI1 increases mInscexpression, an effect that is inhibited by SLIT2 treatment. Increased mInscdoes not change cell proliferation in the mammary gland (MG) but instead causes more basal cap cells to divide via SCD, at the expense of ACD, leading to more stem cells and larger outgrowths. Together, our studies provide insight into how the number of mammary stem cells is regulated by the extracellular cue SLIT2.

An array cold atmospheric pressure plasma jet (ACAPPJ) apparatus with 4 × 4 channels and coverage area over than 6 cm2was developed to decontaminate toxic chemicals on material surfaces using Helium (He) gas as discharge gas. Experimental results indicated that 20.4 mg aniline on 6.25 cm2stainless steel plate could be totally decontaminated within 3 min when the ACAPPJ was conducted under 20 kV, 15 kHz, 10 L/min gas flow and 1.5 cm distance. The decontamination efficiency of malathion was not as good as aniline, but addition of H2O and O2could apparently improve the decontamination performances. The decontamination rate of malathion could be improved from 45% to 95% when 1% H2O and 1% O2were simultaneously introduced with He into the ACAPPJ system. ACAPPJ could even decontaminate the malathion penetrated into wood or rubber materials. The decontamination of aniline and malathion were the first order kinetic reactions, in which the·OH and O3formed in the ACAPPJ played major roles. The decontamination of aniline was mainly electrophilic and ring opening reactions, while the malathion was mainly decontaminated due to the breaking of PS and S-C bonds. The intermediate products of aniline and malathion could be mineralized into CO2and H2O when they were decontaminated by ACAPPJ for enough time. According to the nude mouse skin experiments, the ACAPPJ could be directly applied on animal skins without any electrical and chemical harms.

Air pollution is a risk factor for poor cognitive function, while a plant-based dietary pattern is associated with better cognitive function. We aimed to explore their interaction with cognitive function among older adults.

Our study aimed to test whether the effects of residential greenness on cognitive function and mortality differ by APOEgenotypes among Chinese adults aged 65 years and older.

Both air pollution and green space have been shown to independently affect mortality. We aimed to study whether greenness might interact with air pollution-related mortality among Chinese adults aged 65 years and older.