Artificial Intelligence has taken a notch higher and in its latest marvel, researchers have used Artificial Intelligence technology to estimate obesity on Earth without directing counting the obese people. This modern-day marvel was undertaken by scanning through Google Maps images by a team lead by researchers from Seattle’s University of Washington. The team based their study by downloading nearly 150,000 high-resolution Google Map satellite images of pre-determined localities of four cities, Seattle (Washington State), Los Angeles (California); San Antonio (Texas) and Memphis (Tennessee).
The 500 Cities Project
The leading national public health institute of the United States, Centres for Disease Control and Prevention (CDC) in collaboration with the Robert Wood Johnson Foundation, and the CDC Foundation has taken up the 500 Cities Project to analyze local data for better health. The 500 Cities Project furnishes micro-level estimates for chronic disease risk factors, health outcomes, and clinical preventive service use with the help of data collated from 500 largest cities in the United States. The data analysis model-based estimates for pre-defined health outcomes, prevention and unhealthy behaviors category. The chronic diseases covered are Arthritis, Current Asthma, High Blood Pressure, Cancer (except skin), High Cholesterol, Chronic Kidney disease, COPD, Chronic Heart disease, Diabetes, Mental Health, Physical Health, Teeth Loss and Stroke.
These small area estimates will assist cities, local health departments and urban planners to better understand the geographic distribution of health variables, and assist them to plan public health programmes in a micro level.
Analysing Adult Obesity Prevalence
Data on adult obesity prevalence was obtained from CDC’s “500 Cities” project which appeared in the journal JAMA Network Open. The team lead by researchers from Seattle’s University of Washington fed these images into a convolutional deep learning neural network that extracted features of the built environment, the distribution of buildings and green area.
The analysis of built environment depicted that physical characteristics of a neighborhood like the presence of parks, highways, diverse housing types, green streets, crosswalks etc. can relate to variations in obesity prevalence across different geographical localities. The researchers concluded that Obesity is linked to triggering factors like genetics, diet, physical activity and the environment which cause a positive impact on increasing obesity levels among masses.
Obesity Varies Across Geographical Contexts
Traditionally, there have been evidence which indicate associations between the built environment and obesity vary across studies and geographical boundaries. The study illustrates that convolutional neural networks can be deployed to automate the features extraction of the built environment from satellite images to study the different health indicators. The researchers addressed that it is imperative to understand the association between specific features of the built environment and obesity prevalence to develop structural changes encouraging physical activity leading to a decrease in obesity prevalence.
Sedentary Lifestyle and Behavioural Traits
Behavioral traits encouraging unhealthy food choices and sedentary lifestyles are associated with the social and built environment. The research findings showed that the social and built environments have a deep bearing on the well-being of the health parameters. The built environment can influence health through resource availability like housing, recreational and fitness spaces.
Improving Social Environment and Town Planning
Analysing the micro obesity rates of localities has a great potential to improve the social environment and empowers government authorities and town planners to develop areas specific to fitness spaces like parks, fitness community clubs, jogging open spaces and so on. This provides an immense opportunity to improve the social environment through community health and fitness activities.
Artificial Intelligence and Neural Networks have caused a massive disruption towards a positive change by attempting to identify obesity through the physical characteristics of a neighborhood and built environment. The change has just begun towards a more healthier step to a wholesome living.