Search

Abstract Semantic segmentation plays a crucial role in interpreting remote sensing images, especially in high-resolution scenarios where finer object details, complex spatial information and texture structures exist. To address the challenge of better extracting semantic information and ad-dressing class imbalance in multiclass segmentation, we propose utilizing diffusion models for remote sensing image semantic segmentation, along with a lightweight classification module based on a spatial-channel attention mechanism. Our approach incorporates unsupervised pretrained components with a classification module to accelerate model convergence. The diffusion model component, built on the UNet architecture, effectively captures multiscale features with rich contextual and edge information from images. The lightweight classification module, which leverages spatial-channel attention, focuses more efficiently on spatial-channel regions with significant feature information. We evaluated our approach using three publicly available datasets: Postdam, GID, and Five Billion Pixels. In the test of three datasets, our method achieved the best results. On the GID dataset, the overall accuracy was 96.99%, the mean IoU was 92.17%, and the mean F1 score was 95.83%. In the training phase, our model achieved good performance after only 30 training cycles. Compared with other models, our method reduces the number of parameters, improves the training speed, and has obvious performance advantages.

AbstractChina’s secondary vocational medical education is essential for training primary health care personnel and enhancing public health responses. This education system currently faces challenges, primarily due to its emphasis on knowledge acquisition that overshadows the development and application of skills, especially in the context of emerging artificial intelligence (AI) technologies. This article delves into the impact of AI on medical practices and uses this analysis to suggest reforms for the vocational medical education system in China. AI is found to significantly enhance diagnostic capabilities, therapeutic decision-making, and patient management. However, it also brings about concerns such as potential job losses and necessitates the adaptation of medical professionals to new technologies. Proposed reforms include a greater focus on critical thinking, hands-on experiences, skill development, medical ethics, and integrating humanities and AI into the curriculum. These reforms require ongoing evaluation and sustained research to effectively prepare medical students for future challenges in the field.

Visibility analysis is a crucial geographic information processing method, used to assess the observable spatial range from a particular location under specific conditions. It has broad applications, such as urban planning, landscape design and environmental research. Laser scanning provides detailed and accurate 3D data for visibility analysis, called point cloud. However, certain deficiencies still exist in current point cloud-based visibility analysis methods. Few methods are developed from indoor perspective and the influence of point cloud void is rarely considered. Firstly, indoor viewpoints are generated based on the windows extracted from terrestrial laser scanning (TLS) point cloud and unmanned aerial vehicle (UAV) point cloud is used to provide 3D representation of the environment around the viewpoint. Then, a half-sphere viewshed is created for each indoor viewpoint to analyze the visible spatial range. In this step, a novel line-of-sight correction method is proposed based on the semantic information from UAV point cloud, to correct the biased observation distance resulted from point cloud voids. Lastly, the volume of visible space is calculated based on the corrected viewshed image, to evaluate the visibility of each indoor viewpoint. Our method is validated on two datasets. The results indicate that our method enables more accurate simulation of visible space compared with previous studies by considering the effect of point cloud void. The proposed method can support the urban design from the perspective of indoor visual perception and accurate calculation of other visual factors in visual quality evaluation.

Insulators are key components to ensure the normal operation of power facilities in transmission corridors. Existing insulator identification methods mainly use image data and lack the acquisition of three-dimensional information. This paper proposes an efficient insulator extraction method based on UAV (unmanned aerial vehicle) LiDAR (light detection and ranging) point cloud, using five histogram features: horizontal density (HD), horizontal void (HV), horizontal width (HW), vertical width (VW) and vertical void (VV). Firstly, a voxel-based method is employed to roughly extract power lines and pylons from the original point cloud. Secondly, the VV histogram is used to categorize the pylons into suspension and tension types, and the HD histogram is used to locate the tower crossarm and further refine the roughly extracted powerlines. Then, for the suspension tower, insulators are segmented based on the HV histogram and HD difference histogram. For the tension tower, the HW histogram is used to recognize the jumper conductor (JC) and transmission conductor (TC) from the power line. The HW histogram and VW histogram are used to extract the tension insulator in the TC and suspension insulator in the JC, respectively. Finally, considering the problem of setting a suitable grid width when constructing the feature histogram, an adaptive method of multi-scale histograms is proposed to refine the extraction result. Two 220 kV long transmission lines are used for the validation, and the overall object-based accuracy for suspension and tension towers are 100% and 97.3%, respectively. Compared with the point feature-based method, the mean F1 score of the proposed method improved by 0.3, and the runtime for each tower is within 2 s.

Engineering activities along the Sichuan–Tibet Railway (STR) could cause land degradation and threaten the surrounding ecological security. It is crucial to evaluate the integrated land ecology during and after the construction of this project. This study assesses the land ecology along the STR corridor from 2000 to 2022 using a transfer matrix, the analytic hierarchy process (AHP), and the PSR-TOPSIS model. The main results are as follows: (1) The novel comprehensive ecological assessment process including nine indicators is feasible. (2) The high-quality land ecological, surface vegetation, and environmental regions were concentrated in Ya’an and Nyingchi, whereas the low-quality regions were situated in Qamdo and Garze Tibetan Autonomous Prefecture. (3) There was an overall decline in the integrated land ecological quality along the STR from 2000 to 2022. While it steadily improved in the Ya’an and Nyingchi regions from 2010 to 2022, it continued to decline around the Qamdo region. (4) The most degraded land-use type during the 22 years was grassland, and farmland was the most secure land-use type. Overall, spatial analyses and examinations of residue disposal sites suggested that these locations have negatively impacted integrated land ecology since the beginning of the STR construction project. Our findings have implications for preserving the ecological ecosystem and ensuring the sustainability of the STR construction project.

Accurate detection and extraction of changes in buildings heights is important in monitoring construction (both legal and illegal) and assessing disasters. It is also important information for updating real 3D scenes. However, when using remote sensing images, shadows, vegetation and objects with similar spectral and morphological characteristics as buildings can cause false detections, omissions and incomplete patch edges. To address this issue, we develop the multiscale feature fusion network for dual-modal data (MFFNet), which has two main aspects: (1) The multi-dual-modal feature fusion module detects changes in features with similar spectral and morphological characteristics as buildings. This mitigates false detections by making the model more aware of areas where the elevation has changed over time. (2) Because building extraction is affected by shadows and vegetation, we designed a multiscale feature shuffle module. It takes multiscale features and establishes relationships between neighbouring pixels using the pixel-shuffle algorithm, then fuses and reorganizes the multiscale features to highlight the relationships between global contexts, thereby mitigating the problem of building occlusion by shadows. Comparative experiments show that MFFNet achieves better results on GF7-CD and 3DCD datasets than other similar methods. The proposed method can more accurately monitor building changes over large areas.

Air pollution has become a global concern for many years. Vehicular crowdsensing systems make it possible to monitor air quality at a fine granularity. To better utilize the sensory data with varying credibility, truth discovery frameworks are introduced. However, in urban cities, there is a significant difference in traffic volumes of streets or blocks, which leads to a data sparsity problem for truth discovery. Protecting the privacy of participant vehicles is also a crucial task. We first present a data masking-based privacy-preserving truth discovery framework, which incorporates spatial and temporal correlations to solve the sparsity problem. To further improve the truth discovery performance of the presented framework, an enhanced version is proposed with anonymous communication and data perturbation. Both frameworks are more lightweight than the existing cryptography-based methods. We also evaluate the work with simulations and fully discuss the performance and possible extensions.

Cooking is a very decentralized and private way of energy consumption in human activities. The existing investigation and statistical analysis can’t effectively calculate the proportion of population relying on clean cooking fuel and utilization technology in the region. Therefore, based on the big data of the earth, this paper adopts the combination of spatial analysis and statistical analysis to determine the survey sample area, and according to the economic conditions, topographic characteristics, national policies for new energy construction and living habits of the provinces and cities under investigation, the questionnaire survey is conducted for the research area to calculate the proportion of the regional population relying on clean cooking fuels and utilization technologies. Taking the south of the Yangtze River in China as an example, the paper effectively calculates that 88.25% of the population depends on clean cooking fuel and technology in this region, of which 89.81% are in urban areas and 79.87% are in rural areas. Analysis of the survey data shows that the proportion of the population using clean cooking fuels and technologies is related to factors such as economic development, income and resource endowment. There is a large urban–rural gap in terms of energy consumption tendency and structure in the south of the Yangtze River in China and cooking energy consumption in rural households also varies from region to region. The technical ideas and conclusions of the paper have high reference application value, which can help promote the upgrading of clean energy utilization and provide data basis for relevant decision and policy making.

Highlights 1. COE improves lipid levels and lipase activities in hyperlipidemic rats. 2. COE ameliorates antioxidant enzyme activities and liver function in hyperlipidemic rats. 3. COE exerts a hypolipidemic effect by activating the leptin/JAK2/STAT3 pathway.

Building orientation recognition is a fundamental component of cartographic applications such as map regularization and map generalization. Most current methods effectively calculate orientations for buildings with a single orientation but struggle with complex multi-oriented structures and architectural complexes. This limitation leads to overlooking the spatial relationship between building groups in map generalization. In this paper, we propose a vector reconstruction-based method for calculating building orientation to address the above limitation. By grouping and transforming building edge vectors using a designed vector reconstruction principle, we calculate and correct orientations to obtain the orientation vectors of building footprints. We validate our approach using multiple datasets, and our findings show successful orientation calculation and description for complex buildings and architectural complexes. We also compare our method with the main direction method through map regularization experiments. The results demonstrate that our method better maintains the original shape and geometric structure of buildings, leading to smaller changes in node movement and area. Our approach provides an effective solution for accurately calculating building orientation in complex architectural environments.

The Weather Research and Forecasting (WRF) model was used to simulate Typhoon Rumbia in this paper. The sensitivity experiments were conducted with 16 different parameterization combination schemes, including four microphysics (WSM6, WSM5, Lin, and Thompson), two boundary layers (YSU and MYJ), and two cumulus convection (Kain–Fritsch and Grell–Freitas) schemes. The impacts of 16 parameterization combination schemes and the data assimilation (DA) of Global Navigation Satellite System (GNSS) water vapor were evaluated by the simulation accuracy of typhoon track and intensity. The results show that the typhoon track and intensity are significantly influenced by parameterization schemes of cumulus and boundary layers rather than microphysics. The averaged track error of Lin_KF_Y is 104.73 km in the entire 72-h simulation period. The track errors of all the other combination schemes are higher than Lin_KF_Y. During the entire 72-h, the averaged intensity error of Thompson_GF_M is 1.36 hPa. It is the lowest among all the combination schemes. As for data assimilation, the simulation accuracy of typhoon tracks can be significantly improved by adding the GNSS water vapor. Thompson_GF_M-DA combination scheme has the lowest average track error of 45.05 km in the initial 24 h. The Lin_KF_Y-DA combination scheme exhibits an average track error of 32.17 km on the second day, 28.03 km on the third day, and 35.33 km during 72-h. The study shows that the combination of parameterization schemes and the GNSS water vapor data assimilation significantly improve the initial conditions and the accuracy of typhoon predictions. The study results contribute to the selection of appropriate combinations of physical parameterization schemes for the WRF-ARW model in the mid-latitude region of the western Pacific coast.

Abstract The Gaofen‐7 (GF‐7) satellite is China's first submeter high‐resolution optical stereo‐mapping satellite. The satellite is equipped with two area‐array laser footprint cameras to capture the laser spots emitted by laser altimeters. To accurately identify and locate the geographical positions of laser spots, a stepwise geometric calibration method with ground laboratory information for high‐resolution area‐array camera is proposed in this paper. It includes the internal and external calibration. First, the initial values of the camera's internal parameters are determined based on the laboratory calibration results. Then, based on a high‐precision reference ortho‐image and a digital elevation model, the ground control points (GCPs) for multi‐scene footprint images are obtained by image matching. Finally, the external and internal calibration methods for the laser footprint camera are implemented successively. The experimental results indicate that this method can restore geometric imaging relationships with high precision. After the internal and external calibration, the systematic residual error of the image point is eliminated, the remaining residual achieves 0.38 pixel, and the accuracy of direct positioning on the ground is better than 5.0 m, and the elevation accuracy is within 1.5 m.

Urban vegetation plays a crucial role in the urban ecological system. Efficient and accurate extraction of urban vegetation information has been a pressing task. Although the development of deep learning brings great advantages for vegetation extraction, there are still problems, such as ultra-fine vegetation omissions, heavy computational burden, and unstable model performance. Therefore, a Separable Dense U-Net (SD-UNet) was proposed by introducing dense connections, separable convolutions, batch normalization layers, and Tanh activation function into U-Net. Furthermore, the Fake sample set (NIR-RG), NDVI sample set (NDVI-RG), and True sample set (RGB) were established to train SD-UNet. The obtained models were validated and applied to four scenes (high-density buildings area, cloud and misty conditions area, park, and suburb) and two administrative divisions. The experimental results show that the Fake sample set can effectively improve the model’s vegetation extraction accuracy. The SD-UNet achieves the highest accuracy compared to other methods (U-Net, SegNet, NDVI, RF) on the Fake sample set, whose ACC, IOU, and Recall reached 0.9581, 0.8977, and 0.9577, respectively. It can be concluded that the SD-UNet trained on the Fake sample set not only is beneficial for vegetation extraction but also has better generalization ability and transferability.

Landslides, the second largest geological hazard after earthquakes, result in significant loss of life and property. Extracting landslide information quickly and accurately is the basis of landslide disaster prevention. Fengjie County, Chongqing, China, is a typical landslide-prone area in the Three Gorges Reservoir Area. In this study, we newly integrate Shapley Additive Explanation (SHAP) and Optuna (OPT) hyperparameter tuning into four basic machine learning algorithms: Gradient Boosting Decision Tree (GBDT), Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Additive Boosting (AdaBoost). We construct four new models (SHAP-OPT-GBDT, SHAP-OPT-XGBoost, SHAP-OPT-LightGBM, and SHAP-OPT-AdaBoost) and apply the four new models to landslide extraction for the first time. Firstly, high-resolution remote sensing images were preprocessed, landslide and non-landslide samples were constructed, and an initial feature set with 48 features was built. Secondly, SHAP was used to select features with significant contributions, and the important features were selected. Finally, Optuna, the Bayesian optimization technique, was utilized to automatically select the basic models’ best hyperparameters. The experimental results show that the accuracy (ACC) of these four SHAP-OPT models was above 92% and the training time was less than 1.3 s using mediocre computational hardware. Furthermore, SHAP-OPT-XGBoost achieved the highest accuracy (96.26%). Landslide distribution information in Fengjie County from 2013 to 2020 can be extracted by SHAP-OPT-XGBoost accurately and quickly.

In response to the nonlinear fitting difficulty of the traditional weighted average temperature (Tm) modeling, this paper proposed four machine learning (ML)-based Tm models. Based on the seven radiosondes in the Yangtze River Delta region from 2014 to 2019, four forecasting ML-based Tm models were constructed using Light Gradient Boosting Machine (LightGBM), Support Vector Machine (SVM), Random Forest (RF), and Classification and Regression Tree (CART) algorithms. The surface temperature (Ts), water vapor pressure (Es), and atmospheric pressure (Ps) were identified as crucial influencing factors after analyzing their correlations to the Tm. The ML-based Tm models were trained using seven radiosondes from 2014 to 2018. Then, the mean bias and root mean square error (RMSE) of the 2019 dataset were used to evaluate the accuracy of the ML-based Tm models. Experimental results show that the overall accuracy of the LightGBM-based Tm model is superior to the SVM, CART, and RF-based Tm models under different temporal variations. The mean RMSE of the daily LightGBM-based Tm model is reduced by 0.07 K, 0.04 K, and 0.13 K compared to the other three ML-based models, respectively. The mean RMSE of the monthly LightGBM-based Tm model is reduced by 0.09 K, 0.04 K, and 0.11 K, respectively. The mean RMSE of the quarterly LightGBM-based Tm model is reduced by 0.09 K, 0.04 K, and 0.11 K, respectively. The mean bias of the LightGBM-based Tm model is also smaller than that of the other ML-based Tm models. Therefore, the LightGBM-based Tm model can provide more accurate Tm and is more suitable for obtaining GNSS precipitable water vapor in the Yangtze River Delta region.

The extremely difficult area has the characteristics of extremely rugged terrain, complex geographical environment and sparse permanent scatterers. Therefore, an improved SBAS-InSAR technology is designed to monitor the surface deformation. In this paper, the candidate permanent scatterers are obtained from the coherence, amplitude dispersion index and deformation rate, and then the final permanent scatterers are selected by optical images, which are introduced into the SBAS-InSAR calculation process as orbit refining control points, and finally the surface deformation monitoring in the study area is completed. By comparing and analyzing the conventional PS-InSAR technology and SBAS-InSAR technology, the technology has good application value in extremely difficult areas.

Sentinel-1A data are widely used in interferometric synthetic aperture radar (InSAR) studies due to the free and open access policy. However, the short wavelength (C-band) of Sentinal-1A data leads to decorrelation in numerous applications, especially in vegetated areas. Phase blurring and reduced monitoring accuracy can occur owing to changes in the physical and chemical characteristics of vegetation during the satellite revisit period, which essentially makes poor use of SAR data and increases the time and economic costs for researchers. Interferometric coherence is a commonly used index to measure the interference quality of two single-look complex (SLC) images, and its value can be used to characterize the decorrelation degree. The normalized difference vegetation index (NDVI) is obtained from optical images, and its value can be used to characterize the surface vegetation coverage. In order to solve the problem that Sentinel-1A decorrelation in the vegetated area is difficult to estimate prior to single-look complex interference, this paper selects a vegetated area in Sichuan Province, China as the study area and establishes two two-order linear quantitative models between Landsat8-derived normalized difference vegetation index and Sentinel-1A interferometric coherence in co- and cross-polarization: When NDVI at extremely high and low levels, coherence is close to zero, while NDVI and coherence show two different linear relationships in co- and cross-polarization in terms of NDVI at the middle level. The models global error basically obeys the normal distribution with the mean value of −0.037 and −0.045, and the standard deviation of 0.205 and 0.201 at the VV and VH channels. The two models are then validated in two validation areas, and the results confirm the reliability of the models and reveal the relationships between Sentinel-1A InSAR decorrelation and vegetation coverage in co- and cross-polarization, thus demonstrating that the NDVI can be applied to quantitatively estimate the InSAR decorrelation in vegetated area of Sentinel-1A data in both polarization modes prior to SLC interference.

Ubiquitination is a prevalent post translational modification. Here, the authors show a pivotal role for Nedd4l in the regulation of antiviral immunity via promotion of ubiquitination of TRAF3 and go on to show disruption of Nedd4l both in vitro and in vivo perturbs the antiviral immune response.

Abstract Background Increased fucosylation is associated with the chemoresistance phenotype. Meanwhile, fucosyltransferase IV (FUT4) amounts are frequently elevated in lung cancer and may be related to increased chemoresistance. Methods In the present work, FUT4’s role in cisplatin-induced apoptosis was assessed in A549 and H1975 cells, respectively. To clarify whether the FUT4 gene attenuates chemosensitivity in tumor cells, we constructed FUT4siRNA and evaluated its effects on cisplatin-induced apoptosis and cell growth inhibition. Cell viability, apoptosis, migration and invasion assay were conducted to investigate cisplatin sensitivity. The activation of EGFR/AKT/FOXO1 signaling were measured by western blot. The translocation of FOXO1 was assessed by IFC using Laser Scanning Confocal Microscope. Results We found that FUT4 knockdown dose-dependently increased cisplatin-associated cytotoxicity. Furthermore, FUT4 silencing induced apoptosis and inhibited proliferation in A549 and H1975 cells by suppressing Akt and FOXO1 phosphorylation induced by cisplatin administration, which resulted in nuclear translocation of FOXO1. Conclusion These results suggested FUT4 might control chemoresistance to cisplatin in lung cancer by suppressing FOXO1-induced apoptosis.

Point cloud classification is a key step for three-dimensional (3D) scene analysis in terrestrial laser scanning but is commonly affected by density variation. Many density-adaptive methods are used to weaken the impact of density variation and angular resolution, which denotes the angle between two horizontally or vertically adjacent laser beams and are commonly used as known parameters in those methods. However, it is difficult to avoid the case of unknown angular resolution, which limits the generality of such methods. Focusing on these problems, we propose a density-adaptive feature extraction method, considering the case when the angular resolution is unknown. Firstly, we present a method for angular resolution estimation called neighborhood analysis of randomly picked points (NARP). In NARP, n points are randomly picked from the original data and the k nearest points of each point are searched to form the neighborhood. The angles between the beams of each picked point and its corresponding neighboring points are used to construct a histogram, and the angular resolution is calculated by finding the adjacent beams of each picked point under this histogram. Then, a grid feature called relative projection density is proposed to weaken the effect of density variation based on the estimated angular resolution. Finally, a 12-dimensional feature vector is constructed by combining relative projection density and other commonly used geometric features, and the semantic label is generated utilizing a Random Forest classifier. Five datasets with a known angular resolution are used to validate the NARP method and an urban scene with a scanning distance of up to 1 km is used to compare the relative projection density with traditional projection density. The results demonstrate that our method achieves an estimation error of less than 0.001° in most cases and is stable with respect to different types of targets and parameter settings. Compared with traditional projection density, the proposed relative projection density can improve the performance of classification, particularly for small-size objects, such as cars, poles, and scanning artifacts.

Extreme loadings, a hostile environment and dangerous operation lead to the unsafe state of bridges under construction, especially large-span bridges. Global Navigation Satellite Systems (GNSS) tend to be the best choice for real-time deformation monitoring due to the significant advantage of automation, continuation, all-weather operation and high precision. Unfortunately, the traditional geodetic GNSS instrument with its high price and large volume is limited in its applications. Hence, we design and develop low-cost GNSS equipment by simplifying the monitoring module. The performance of the proposed solution is evaluated through an experimental dynamic scenario, proving its ability to track abrupt deformation down to 3–5 mm. We take Chongqing Guojiatuo Suspension Bridge in China as a case study. We build a real-time low-cost GNSS monitoring cloud platform. The low-cost bridge GNSS monitoring stations are located at the top of the south and north towers, midspan upstream and downstream respectively and the reference station is located in the stable zone 400 m away from the bridge management buildings. We conducted a detailed experimental assessment of low-cost GNSS on 5 April and a real-time deformation detection experiment of the towers and main cables during the dynamic cable saddle pushing process on 26 February 2022. In the static experiment, the standard deviation of the residual using the multi-GNSS solution is 2 mm in the horizontal direction and 5 mm in the vertical direction. The multi-GNSS solution significantly outperforms the BDS/GPS single system. The dynamic experiment shows that, compared with the movement measured by the robotic total station, the horizontal error of the south tower and north tower measured by low-cost GNSS is below 0.005 m and 0.008 m respectively. This study highlights the potential of low-cost GNSS solutions for Structural Health Monitoring (SHM) applications.

TcpC is a multifunctional virulence factor of Uropathogenic Escherichia coli (UPEC). Macrophages can differentiate into two different subsets M1 and M2 that play distinct roles in anti-infection immunity. Here, we investigate the influence of TcpC on M1/M2 polarization and the potential mechanisms. Our data showed that M1 markers CD86 and iNOS were significantly inhibited, while the M2 markers CD163, CD206 and Arg-1 were enhanced in macrophages in kidneys from the TcpC-secreting wild-type CFT073 (CFT073wt)-infected pyelonephritis mouse model, compared with those in macrophages in kidneys from TcpC knockout CFT073 mutant (CFT073Δtcpc)-infected mice. CFT073wt or recombinant TcpC (rTcpC) treatment inhibits LPS + IFN-γ-induced CD80, CD86, TNF-α and iNOS expression, but promotes IL-4-induced CD163, CD206, Arg-1 and IL-10 expression in both human and mouse macrophage cell lines THP-1 and J774A.1. Moreover, rTcpC significantly attenuated LPS + IFN-γ-induced phosphorylation of p38, ERK, p50 and p65 but enhanced IL-4-induced phosphorylation of Akt and STAT6. These data suggest that TcpC inhibits M1 but promotes M2 macrophage polarization by down-regulation of p38, ERK/NF-κB and up-regulation of the Akt/STAT6 signaling pathway, respectively. Our findings not only illuminate the regulatory effects of TcpC on macrophage M1/M2 polarization and its related signaling pathways, but also provide a novel mechanism underlying TcpC-mediated immune evasion of macrophage-mediated innate immunity.

Nowadays, large-scale human mobility has led to increasingly severe traffic congestion in cities, how to accurately identify people’s travel mode has become particularly important for urban traffic planning and management. However, traditional methods are based on telephone interviews or questionnaires, which makes it difficult to obtain accurate and effective data. Nowadays, numbers of smartphones are equipped with various sensors, including accelerometers, gyroscopes, and GPS, providing a novel social sensing data source to detect people’s travel modes. The fusion of multiple sensor data is a promising way for travel mode detection. However, how to use these sensor data to accurately detect travel mode is still a challenging task. In this paper, we presented a light-weight method for travel mode detection based on four types of smartphone sensor data collected from an accelerometer, gyroscope, magnetometer, and barometer, and a prototype application was developed. Then, a novel convolutional neural network (CNN) was designed to identify five representative travel modes (walk, bicycle, bus, car, and metro). We compared the overall performance of the proposed method via different hyperparameters, and the experimental results show that the F value of the proposed method reaches 97%, which verified the effectiveness of the proposed method for travel mode classification.

Through utilizing user mobility and short-range device-to-device communication techniques, mobile opportunistic networks (MONs) enable end-to-end message delivery without the dependence on reliable network infrastructures. Examples of these networks include mobile social networks and vehicular networks. Multicast in MONs is used to disseminate information to a group of mobile nodes, which have attracted considerable attention due to high resource utilization. In this paper, we focus on reducing the overhead of multicast in MONs without compromising the delivery performance, through utilizing static social features of nodes and time-varying social behaviors. We first conduct a trace data analysis using the information entropy theory to identify the most important and representative social features in a popular trace, the Infocom06 trace. Based on these static social features, we propose a Social Profile-based Multicast (SPM) routing scheme, that supports efficient multicast message dissemination with a small maintenance overhead, i.e., little cost on maintaining the historical records. Furthermore, by exploring the time-varying social behaviors during daytime and nighttime, we verify that a small number of forwardings during that daytime is sufficient to achieve a desirable delivery ratio. We thus propose an improved overhead-reducing scheme Social Profile-based Multicast-Overhead Reducing (SPMOR) that restricts the number of forwardings during the daytime. The extensive trace-driven simulations show that SPM achieves desirable delivery performance with small maintenance and transmission cost, and SPMOR can further reduce the transmission overhead under diurnal user behaviors. At last, we conduct a similar study on a campus-based MON trace, SocialBlueConn. We find that the main conclusions and performance results on SocialBlueConn are consistent with the Infocom06 trace, which verifies our methodology is scalable.

As a local medicine and food, wampee fruit, with abundant bioactive compounds, is loved by local residents in Southern China. Titratable acid (TA), total sugar (TS), and total phenolic and flavonoid contents were detected, and phytochemical profiles and cellular antioxidant activities were analyzed by the HPLC and CAA (cellular antioxidant activity) assay in five sweet wampee varieties and five sour wampee varieties. Results showed that the average TS/TA ratio of sweet wampee varieties was 29 times higher than sour wampee varieties, while TA content was 19 times lower than sour wampee varieties. There were much lower levels of total phenolics, flavonoids, and antioxidant activities in sweet wampee varieties than those in sour wampee varieties. Eight phytochemicals were detected in sour wampee varieties, including syringin, rutin, benzoic acid, 2-methoxycinnamic acid, kaempferol, hesperetin, nobiletin, and tangeretin, while just four of them were detected in sweet wampee varieties. Syringin was the only one that was detected in all the sour wampee varieties and was not detected in all sweet wampee varieties. Correlation analysis showed significant positive correlations between TA with phenolics, flavonoids, and total and cellular (PBS wash) antioxidant activities, while there were significant negative correlations between TS/TA with phenolic and cellular (no PBS wash) antioxidant activities. This suggested that the content of titratable acid in wampee fruit might have some relationship with the contents of phenolics and flavonoids. Sour wampee varieties should be paid much attention by breeders for their high phytochemical contents and antioxidant activities for cultivating germplasms with high health care efficacy.

Real-time monitoring of urban building development provides a basis for urban planning and management. Remote sensing change detection is a key technology for achieving this goal. Intelligent change detection based on deep learning of remote sensing images is a current focus of research. However, most methods only use unimodal remote sensing data and ignore vertical features, leading to incomplete characterization, poor detection of small targets, and false detections and omissions. To solve these problems, we propose a multi-path self-attentive hybrid coding network model (MAHNet) that fuses high-resolution remote sensing images and digital surface models (DSMs) for 3D change detection of urban buildings. We use stereo images from the Gaofen-7 (GF-7) stereo mapping satellite as the data source. In the encoding stage, we propose a multi-path hybrid encoder, which is a structure that can efficiently perform multi-dimensional feature mining of multimodal data. In the deep feature fusion link, a dual self-attentive fusion structure is designed that can improve the deep feature fusion and characterization of multimodal data. In the decoding stage, a dense skip-connection decoder is designed that can fuse multi-scale features flexibly and reduce spatial information losses in small-change regions in the down-sampling process, while enhancing feature utilization and propagation efficiency. Experimental results show that MAHNet achieves accurate pixel-level change detection in complex urban scenes with an overall accuracy of 97.44% and F1-score of 92.59%, thereby outperforming other methods of change detection.