Air quality improvement assessment and exposure risk of Shandong Province in China during 2014 to 2020.

As one of the most polluted provinces in China, air pollution events occur frequently in Shandong. Based on the hourly (or daily) concentrations of six air pollutants (PM2.5, PM10, O3, NO2, SO2 and CO), the situations of air quality improvement in three kinds of cities (key cities, coastal cities and general cities) are assessed comprehensively during 2014-2020. Contrary to the daily maximum 8-h average ozone (MDA8 O3), the annual average concentrations of other pollutants show the downward trends during 2014-2020. Therein, the improvement rates of annual average concentrations of air pollutants in key cities are highest. By 2020, the day proportions of O3 as the primary pollutant are up to 38% in three kinds of cities. Besides, due to the impact of COVID-19, the monthly average concentrations of PM2.5, PM10, NO2, SO2 and CO in February 2020 decrease by 32.1-49.5% year-on-year. There are still about 50% of population exposed to high-risk regions (R i > 2), which are mainly concentrated in main urban areas and industrial areas. Thus, the adjustment of industrial structure and energy composition in the context of carbon peak and carbon neutrality should be implemented in the future. Supplementary Information: The online version contains supplementary material available at 10.1007/s13762-022-04651-5.

An Effective Deep Neural Network for Lung Lesions Segmentation From COVID-19 CT Images

Automatic segmentation of lung lesions from COVID-19 computed tomography (CT) images can help to establish a quantitative model for diagnosis and treatment. For this reason, this article provides a new segmentation method to meet the needs of CT images processing under COVID-19 epidemic. The main steps are as follows: First, the proposed region of interest extraction implements patch mechanism strategy to satisfy the applicability of 3-D network and remove irrelevant background. Second, 3-D network is established to extract spatial features, where 3-D attention model promotes network to enhance target area. Then, to improve the convergence of network, a combination loss function is introduced to lead gradient optimization and training direction. Finally, data augmentation and conditional random field are applied to realize data resampling and binary segmentation. This method was assessed with some comparative experiment. By comparison, the proposed method reached the highest performance. Therefore, it has potential clinical applications.