Research on damage detection of road surfaces using image processing techniques has been actively conducted. This study makes three contributions to address road damage detection issues. First, to the best of our knowledge, for the first time, a large‐scale road damage data set is prepared, comprising 9,053 road damage images captured using a smartphone installed on a car, with 15,435 instances of road surface damage included in these road images. Next, we used state‐of‐the‐art object detection methods using convolutional neural networks to train the damage detection model with our data set, and compared the accuracy and runtime speed on both, using a GPU server and a smartphone. Finally, we demonstrate that the type of damage can be classified into eight types with high accuracy by applying the proposed object detection method. The road damage data set, our experimental results, and the developed smartphone application used in this study are publicly available (https://github.com/sekilab/RoadDamageDetector/).

Machine learning can produce promising results when sufficient training data are available; however, infrastructure inspections typically do not provide sufficient training data for road damage. Given the differences in the environment, the type of road damage and the degree of its progress can vary from structure to structure. The use of generative models, such as a generative adversarial network (GAN) or a variational autoencoder, makes it possible to generate a pseudoimage that cannot be distinguished from a real one. Combining a progressive growing GAN along with Poisson blending artificially generates road damage images that can be used as new training data to improve the accuracy of road damage detection. The addition of a synthesized road damage image to the training data improves the F-measure by 5% and 2% when the number of original images is small and relatively large, respectively.

AimLittle is known about whether and how local-level resources regarding home care are associated with the prevalence of home deaths. We aimed to investigate whether geographic patterns of the resources for home care were associated with the prevalence of home deaths, taking spatial variation into consideration.MethodsWe conducted an ecological cross-sectional study in Japan using nationwide data in 2014. The areal unit was the municipality, the smallest administrative unit in Japan. We investigated the association between the percentage of home deaths and the resources of home care support clinics with available 24-hour-a-day functions, considering the geographic effect of neighboring municipalities by applying a geographically weighted regression model.ResultsThe mean and standard deviation of the percentages of home deaths were 11.4% (5.0%), and those of the number of home care support clinics per 10,000 elderly population were 3.4 (3.7). The percentages of home deaths in neighboring municipalities tended to be significantly correlated (Moran’s I 0.34, p<0.001). Adjusting for the number of hospital beds, total population, and the socio-economic status of municipality, the results of an ordinary least squares regression model showed a positive correlation between the percentage of home deaths and the local resources for home care support clinics per 10,000 elderly population (regression coefficient 0.15, 95% confidence interval 0.07, 0.22), while the existence of spatial autocorrelation of the residual was suggested (Moran’s I of the residual 0.227, p<0.001). The geographically weighted regression model showed local regression coefficients varying across municipalities with a better model fit over the analogous ordinary least squares model (adjusted R2 0.414 vs. 0.131).ConclusionHome deaths were more prevalent in municipalities with greater home care resources. This association was geographically varied and further strengthened in some areas.

BackgroundIn Japan, the revision of the fee schedules in 2006 introduced a new category of general care ward for more advanced care, with a higher staffing standard, a patient-to-nurse ratio of 7:1. Previous studies have suggested that these changes worsened inequalities in the geographic distribution of nurses, but there have been few quantitative studies evaluating this effect. This study aimed to investigate the association between the distribution of 7:1 beds and the geographic distribution of hospital nursing staffs.MethodsWe conducted a secondary data analysis of hospital reimbursement reports in 2012 in Japan. The study units were secondary medical areas (SMAs) in Japan, which are roughly comparable to hospital service areas in the United States. The outcome variable was the nurse density per 100,000 population in each SMA. The 7:1 bed density per 100,000 population was the main independent variable. To investigate the association between the nurse density and 7:1 bed density, adjusting for other variables, we applied a multiple linear regression model, with nurse density as an outcome variable, and the bed densities by functional category of inpatient ward as independent variables, adding other variables related to socio-economic status and nurse workforce. To investigate whether 7:1 bed density made the largest contribution to the nurse density, compared to other bed densities, we estimated the standardized regression coefficients.ResultsThere were 344 SMAs in the study period, of which 343 were used because of data availability. There were approximately 553,600 full time equivalent nurses working in inpatient wards in hospitals. The mean (standard deviation) of the full time equivalent nurse density was 426.4 (147.5) and for 7:1 bed density, the figures were 271.9 (185.9). The 7:1 bed density ranged from 0.0 to 1,295.5. After adjusting for the possible confounders, there were more hospital nurses in the areas with higher densities of 7:1 beds (standardized regression coefficient 0.62, 95% confidence interval 0.56–0.68).ConclusionWe found that the 7:1 nurse staffing standard made the largest contribution to the geographic distribution of hospital nurses, adjusted for socio-economic status and nurse workforce-related factors.Electronic supplementary materialThe online version of this article (doi:10.1186/s12912-017-0219-1) contains supplementary material, which is available to authorized users.

No abstract