Deep CNNs with Robust LBP Guiding Pooling for Face Recognition.

Pooling layer in Convolutional Neural Networks (CNNs) is designed to reduce dimensions and computational complexity. Unfortunately, CNN is easily disturbed by noise in images when extracting features from input images. The traditional pooling layer directly samples the input feature maps without considering whether they are affected by noise, which brings about accumulated noise in the subsequent feature maps as well as undesirable network outputs. To address this issue, a robust Local Binary Pattern (LBP) Guiding Pooling (G-RLBP) mechanism is proposed in this paper to down sample the input feature maps and lower the noise impact simultaneously. The proposed G-RLBP method calculates the weighted average of all pixels in the sliding window of this pooling layer as the final results based on their corresponding probabilities of being affected by noise, thus lowers the noise impact from input images at the first several layers of the CNNs. The experimental results show that the carefully designed G-RLBP layer can successfully lower the noise impact and improve the recognition rates of the CNN models over the traditional pooling layer. The performance gain of the G-RLBP is quite remarkable when the images are severely affected by noise.

Unveiling the hidden impact: How human disturbances threaten aquatic microorganisms in cities.

Urban activity emissions have important ecological significance to bacterial communities' spatial and temporal distribution and the mechanism of bacterial community construction. The mechanism of bacterial community construction is the key to community structure and lifestyle, and the influence of this aspect has not been thoroughly studied. This study analyzed the response of bacteria in water and sediment in different seasons to urban activities in Jinsha River. The results showed that the influence of urban activities on bacterial community structure in sediment was greater than that in water. The input of pollution in different regions changed the diversity and abundance of water and sediments bacteria and promoted bacterial community reconstruction to a certain extent. Co-network analysis found that many metal-mediated species are core species within the same module and can be used to mitigate pollution caused by metal or organic pollutants due to interspecific solid interactions. Different potential pollution sources around urban rivers affect the metabolic function of bacteria in aquatic ecosystems and promote the detoxification function of bacteria in different media. The results of this study supplement our understanding of the characteristics of microbial communities in urban river systems and provide clues for understanding the maintenance mechanism of microbial diversity in multi-pollution environments.

Study on Radiation Shielding Properties of New Barium-Doped Zinc Tellurite Glass.

This study aimed to investigate the effect of BaF2 on the radiation-shielding ability of lead telluride glass. A physical radioactive source was used to estimate the mass attenuation coefficient (µm) of the 60TeO2-20PbO-(20-x)ZnO-xBaF2 glass system (where x = 1,2,3,5,6,7,9 mol%). We tested the µm values at seven energies (0.059, 0.081, 0.122, 0.356, 0.662, 1.173, 1.332 MeV). To determine the accuracy of the obtained results, we compared the experimental data with the data calculated using the XCOM software. The experimental values obtained for the selected lead telluride glasses at different concentrations of BaF2 are in good agreement with the results of XCOM at all energies. The addition of BaF2 increased the µm value of the sample. At the same time, the half-value layer (HVL), mean free path (MFP), effective atomic number (Zeff), and fast neutron removal cross-section (RCS) of the glass were studied. With the increase in the BaF2 content, the HVL value and MFP value of the glass decreased, and the Zeff value and RCS increased, indicating that the addition of BaF2 enhanced the radiation-shielding performance of the glass.

Predicting soil cadmium uptake by plants in a tailings reservoir during 48-year vegetation restoration.

Plant uptake can reduce soil cadmium (Cd) pollution, while how to exactly predict plant Cd uptake in industrial or mining areas during vegetation restoration remains unexplored. Taking Heteropogon contortus as the object plant, we predicted plant Cd uptake in the Majiatian tailings reservoir during 48-year vegetation restoration by the methods of soil total Cd, DGT (diffusive gradients in thin films technique) and acetic acid (HAc) extraction. Meanwhile, we explored the effects of soil properties on the accuracy of the prediction. Total Cd concentrations in the soils exhibited a better prediction of plant Cd uptake relative to the methods of HAc extraction and DGT. However, the DGT method effectively predicted plant Cd uptake at low Cd supply (lower than 0.42 µg/L), probably because of the dominant diffusion limitation by plants. The prediction of plant Cd uptake by HAc extraction was improved when combined with soil pH. Our results indicate that with increasing external Cd inputs during the vegetation restoration, soil total Cd and traditional extraction method in combination with soil properties are effective ways to predict plant Cd uptake, especially when the Cd fractions cannot be measured by DGT. However, the DGT method works once plant Cd uptake dominated by diffusion limitation despite the interference in soil properties.

STUDY ON THE SHIELDING AND DOSE RATE DISTRIBUTIONS OF THERAPEUTIC PROTON SYNCHROTRON ACCELERATOR BASED ON FLUKA.

Some therapeutic proton synchrotron accelerators will be built for cancer treatment in China. The radiation produced by proton must be carefully evaluated and shielded for occupational disease hazard assessment and environmental impact assessment. Adopting the FLUKA code, a therapeutic room model, a synchrotron hall model and a high energy transport line tunnel model are constructed to get the ambient dose equivalent rate distributions. The ambient dose equivalent rates are also calculated with an empirical formula in some regions. The results calculated by the two ways are consistent with each other. The advantages and disadvantages of the two methods have been analyzed. The ambient dose equivalent rates are no more than 2.5 µSv/h at 30 cm beyond the shielding wall and the maze door. The dose rates are <25 µSv/h at 30 cm beyond the roof of the therapeutic room. These results comply with the National Occupational Health Standard requirements.