Smart Eyes on Screens: AI System Counts Food Ads

Researchers designed a clever computer program called SCANNER Food to monitor how food and beverage companies market themselves across digital spaces. The project focused on building software that can automatically recognize and track branded advertisements in photos and videos posted online. This effort aims to give health authorities fresh data about promotional tactics aimed at shoppers, particularly young people who are heavily influenced by digital content.

The development process involved two major steps spread across 2024 and 2025. Initially, researchers measured how well the algorithm performed using standard machine learning benchmarks. Later, they applied the system to actual social media screen recordings and live video feeds. In controlled testing, the model reached a mean average precision of 0.95, a recall value of 0.96, and an F1 score of 0.96. When evaluated in real environments, it identified 88.5 percent of genuine ads while producing only 11.2 percent of false positives.

These impressive results demonstrate that AI can effectively handle the task of spotting food marketing in visual media. Such capabilities could later enable government agencies to assess compliance with advertising laws, compare spending patterns among brands, and design better strategies to shield vulnerable populations. Ultimately, the proof-of-concept suggests that intelligent automation can serve as a helpful partner in public health protection without removing human oversight.

actions