Training Computers to See the Built Environment Related to Physical Activity: Detection of Microscale Walkability Features Using Computer Vision

The mitigating effect of street trees, urban flora, and the suburban environment on seasonal peak UV indices: A case study from Brisbane, Australia

Tree shade, particularly shade that obscures direct sunlight near peak periods of midday solar exposure can have a pronounced effect on potentially harmful ultraviolet radiation, and in turn, strongly influence the maximum daily UV index (UVI). In this study, the seasonal influence of tree shade on the UVI is evaluated from 210 hemispherical sky view images collected alongside public walkways and footpaths from 10 residential Brisbane suburbs. The effective sidewalk UV index is calculated underneath planted tree canopies, adjacent residential gardens, buildings and background tree species. Results are presented with respect to seasonal variations in the diurnal solar elevation for each month of the year at Brisbane's latitude. The research also examines the total reduction in UVI due to the presence of individual tree species, showing reductions in the midday UVI of up to 91% of an equivalent unimpeded sky hemisphere when overhead tree canopies are present. Important footpath tree species for peak midday UVI mitigation include Pongamia pinnata, Xanthostemon chrysanthus, Senna siamea, and Libidibia ferrea. The planting and maintenance of existing tree species already growing alongside residential Brisbane streets will improve the shade characteristics of suburbs and enhance UV protection for local residents.

Multi-Task Classification for Improved Health Outcome Prediction Based on Environmental Indicators

This paper aims to address the challenges associated with evaluating the impact of neighborhood environments on health outcomes. Google street view (GSV) images provide a valuable tool for assessing neighborhood environments on a large scale. By annotating the GSV images with labels indicating the presence or absence of specific neighborhood features, we can develop classifiers capable of automatically analyzing and evaluating the environment. However, the process of labeling GSV images to analyze and evaluate the environment is a time-consuming and labor-intensive task. To overcome these challenges, we propose using a multi-task classifier to enhance the training of classifiers with limited supervised GSV data. Our multi-task classifier utilizes readily available, inexpensive online images collected from Flickr as a related classification task. The hypothesis is that a classifier trained on multiple related tasks is less likely to overfit to small amounts of training data and generalizes better to unseen data. We leverage the power of multiple related tasks to improve the classifier's overall performance and generalization capability. Here we show that, with the proposed learning paradigm, predicted labels for GSV test images are more accurate. Across different environment indicators, the accuracy, F1 score and balanced accuracy increase up to 6 % in the multi-task learning framework compared to its single-task learning counterpart. The enhanced accuracy of the predicted labels obtained through the multi-task classifier contributes to a more reliable and precise regression analysis determining the correlation between predicted built environment indicators and health outcomes. The R2 values calculated for different health outcomes improve by up to 4 % using multi-task learning detected indicators.

Neighborhood Socioeconomic Status, Green Space, and Walkability and Risk for Falls Among Postmenopausal Women: The Women's Health Initiative

PURPOSE

This study estimated associations between neighborhood socioeconomic status (NSES), walkability, green space, and incident falls among postmenopausal women and evaluated modifiers of these associations, including study arm, race and ethnicity, baseline household income, baseline walking, age at enrollment, baseline low physical functioning, baseline fall history, climate region, and urban-rural residence.

METHODS

The Women's Health Initiative recruited a national sample of postmenopausal women (50-79 years) across 40 U.S. clinical centers and conducted yearly assessments from 1993 to 2005 (n = 161,808). Women reporting a history of hip fracture or walking limitations were excluded, yielding a final sample of 157,583 participants. Falling was reported annually. NSES (income/wealth, education, occupation), walkability (population density, diversity of land cover, nearby high-traffic roadways), and green space (exposure to vegetation) were calculated annually and categorized into tertiles (low, intermediate, high). Generalized estimating equations assessed longitudinal relationships.

RESULTS

NSES was associated with falling before adjustment (high vs. low, odds ratio, 1.01; 95% confidence interval, 1.00-1.01). Walkability was significantly associated with falls after adjustment (high vs. low, odds ratio, 0.99; 95% confidence interval, 0.98-0.99). Green space was not associated with falling before or after adjustment. Study arm, race and ethnicity, household income, age, low physical functioning, fall history, and climate region modified the relationship between NSES and falling. Race and ethnicity, age, fall history, and climate region modified relationships between walkability and green space and falling.

CONCLUSIONS

Our results did not show strong associations of NSES, walkability, or green space with falling. Future research should incorporate granular environmental measures that may directly relate to physical activity and outdoor engagement.

Using Convolutional Neural Networks to Derive Neighborhood Built Environments from Google Street View Images and Examine Their Associations with Health Outcomes

Built environment neighborhood characteristics are difficult to measure and assess on a large scale. Consequently, there is a lack of sufficient data that can help us investigate neighborhood characteristics as structural determinants of health on a national level. The objective of this study is to utilize publicly available Google Street View images as a data source for characterizing built environments and to examine the influence of built environments on chronic diseases and health behaviors in the United States. Data were collected by processing 164 million Google Street View images from November 2019 across the United States. Convolutional Neural Networks, a class of multi-layer deep neural networks, were used to extract features of the built environment. Validation analyses found accuracies of 82% or higher across neighborhood characteristics. In regression analyses controlling for census tract sociodemographics, we find that single-lane roads (an indicator of lower urban development) were linked with chronic conditions and worse mental health. Walkability and urbanicity indicators such as crosswalks, sidewalks, and two or more cars were associated with better health, including reduction in depression, obesity, high blood pressure, and high cholesterol. Street signs and streetlights were also found to be associated with decreased chronic conditions. Chain link fence (physical disorder indicator) was generally associated with poorer mental health. Living in neighborhoods with a built environment that supports social interaction and physical activity can lead to positive health outcomes. Computer vision models using manually annotated Google Street View images as a training dataset were able to accurately identify neighborhood built environment characteristics. These methods increases the feasibility, scale, and efficiency of neighborhood studies on health.