Visual Recognition of Traffic Signs in Natural Scenes Based on Improved RetinaNet.

Aiming at recognizing small proportion, blurred and complex traffic sign in natural scenes, a traffic sign detection method based on RetinaNet-NeXt is proposed. First, to ensure the quality of dataset, the data were cleaned and enhanced to denoise. Secondly, a novel backbone network ResNeXt was employed to improve the detection accuracy and effection of RetinaNet. Finally, transfer learning and group normalization were adopted to accelerate our network training. Experimental results show that the precision, recall and mAP of our method, compared with the original RetinaNet, are improved by 9.08%, 9.09% and 7.32%, respectively. Our method can be effectively applied to traffic sign detection.

MRL-Net: Multi-Scale Representation Learning Network for COVID-19 Lung CT Image Segmentation.

Accuracy segmentation of COVID-19 lesions in lung CT images can aid patient screening and diagnosis. However, the blurred, inconsistent shape and location of the lesion area poses a great challenge to this vision task. To tackle this issue, we propose a multi-scale representation learning network (MRL-Net) that integrates CNN with Transformer via two bridge unit: Dual Multi-interaction Attention (DMA) and Dual Boundary Attention (DBA). First, to obtain multi-scale local detailed feature and global contextual information, we combine low-level geometric information and high-level semantic features extracted by CNN and Transformer, respectively. Secondly, for enhanced feature representation, DMA is proposed to fuse the local detailed feature of CNN and the global context information of Transformer. Finally, DBA makes our network focus on the boundary features of the lesion, further enhancing the representational learning. Amounts of experimental results show that MRL-Net is superior to current state-of-the-art methods and achieves better COVID-19 image segmentation performance.

Molecular characterization of invasive Streptococcus pneumoniae clinical isolates from a tertiary children's hospital in eastern China.

Streptococcus pneumoniae is a common opportunistic pathogen that causes invasive pneumococcal disease (IPD), especially in children. This study aimed to determine the prevalence and molecular characteristics of S. pneumoniae isolated from children with IPD. A total of 78 S. pneumoniae isolates from aseptic body fluids of 70 IPD patients were collected at the Children's Hospital of Nanjing Medical University (Jiangsu Province, China) during 2017-2021. Whole-genome sequencing technology was used to analyze the serotype, sequence type (ST), virulence, and antibiotic resistance of the 78 invasive S. pneumoniae clinical isolates. Our results showed that the pneumococcal infection rate declined after the COVID-19 outbreak in 2019. Serotypes 19F, 14, 6A, 23F, 19A, and 6B were the most common strains. The pneumococcal conjugate vaccine (PCV) 13 serotype coverage rate was 87.1%. All isolates were classified by multi-locus sequence typing (MLST) analysis into 27 different STs, including 3 novel STs (ST17941, ST17942, and ST17944) and 1 novel allele [recP (558)]. The most predominant ST was ST271, followed by ST320 and ST876. All isolates carried the following virulence genes: cbpG, lytB, lytC, pce (cbpE), pavA, slrA, plr (gapA), hysA, nanA, eno, piuA, psaA, cppA, iga, htrA (degP), tig (ropA), zmpB, and ply. All isolates were multidrug resistant and had high levels of resistance to macrolides, tetracyclines, and sulfonamides. Taken together, this study revealed extensive genetic diversity among S. pneumoniae isolates from a single Chinese hospital. Wearing masks, universal infant vaccination with PCV13, and the launch of recombinant protein vaccine development programs could reduce the burden of IPD in children. IMPORTANCE Invasive pneumococcal disease (IPD) caused by Streptococcus pneumoniae in children remains a global burden and should be given more attention due to the fact that the pneumococcal vaccine is not fully covered globally. The molecular epidemiological characteristics of S. pneumoniae are not so clear, especially in these years of COVID-19. In this study, we collected S. pneumoniae isolates from the aseptic body fluid of children with IPD from 2017 to 2021 in a tertiary children's hospital in China and revealed the extensive genetic diversity of these isolates. Most importantly, we first found that the rate of pneumococcal infection has declined since the COVID-19 outbreak in 2019, which means that wearing masks could reduce the transmission of S. pneumoniae. In addition, it was shown that universal infant vaccination with PCV13 seems essential for reducing the burden of IPD in children.

COVID-19 CT image segmentation based on improved Res2Net.

PURPOSE: Corona virus disease 2019 (COVID-19) is threatening the health of the global people and bringing great losses to our economy and society. However, computed tomography (CT) image segmentation can make clinicians quickly identify the COVID-19-infected regions. Accurate segmentation infection area of COVID-19 can contribute screen confirmed cases. METHODS: We designed a segmentation network for COVID-19-infected regions in CT images. To begin with, multilayered features were extracted by the backbone network of Res2Net. Subsequently, edge features of the infected regions in the low-level feature f2 were extracted by the edge attention module. Second, we carefully designed the structure of the attention position module (APM) to extract high-level feature f5 and detect infected regions. Finally, we proposed a context exploration module consisting of two parallel explore blocks, which can remove some false positives and false negatives to reach more accurate segmentation results. RESULTS: Experimental results show that, on the public COVID-19 dataset, the Dice, sensitivity, specificity, S α ${S}_\alpha $ , E ∅ m e a n $E_\emptyset ^{mean}$ , and mean absolute error (MAE) of our method are 0.755, 0.751, 0.959, 0.795, 0.919, and 0.060, respectively. Compared with the latest COVID-19 segmentation model Inf-Net, the Dice similarity coefficient of our model has increased by 7.3%; the sensitivity (Sen) has increased by 5.9%. On contrary, the MAE has dropped by 2.2%. CONCLUSIONS: Our method performs well on COVID-19 CT image segmentation. We also find that our method is so portable that can be suitable for various current popular networks. In a word, our method can help screen people infected with COVID-19 effectively and save the labor power of clinicians and radiologists.

COVID-19 diagnosis via chest X-ray image classification based on multiscale class residual attention.

Aiming at detecting COVID-19 effectively, a multiscale class residual attention (MCRA) network is proposed via chest X-ray (CXR) image classification. First, to overcome the data shortage and improve the robustness of our network, a pixel-level image mixing of local regions was introduced to achieve data augmentation and reduce noise. Secondly, multi-scale fusion strategy was adopted to extract global contextual information at different scales and enhance semantic representation. Last but not least, class residual attention was employed to generate spatial attention for each class, which can avoid inter-class interference and enhance related features to further improve the COVID-19 detection. Experimental results show that our network achieves superior diagnostic performance on COVIDx dataset, and its accuracy, PPV, sensitivity, specificity and F1-score are 97.71%, 96.76%, 96.56%, 98.96% and 96.64%, respectively; moreover, the heat maps can endow our deep model with somewhat interpretability.