Face Recognition for Crowd Analytics in Digital Signage CMS

Two international digital signage providers reached out to 3DiVi team, eager to integrate computer vision algorithms to transform their advertising strategies. They wanted to gather detailed audience statistics, including:

• The number of unique viewers who engaged with the ads

• The average attention time

• The count of people who walked past but didn’t stop to view the ads

• Demographic insights such as age, gender, and emotional state of ad audience

With this data, display owners could refine their ad targeting, optimize screen placement, and maximize engagement—all driving better revenue results.

What is Digital Signage?

Digital signage refers to digital screens used in public spaces to display information, ads, and multimedia content. These screens, commonly placed in locations like shopping malls, airports, and transit hubs, are powered by Digital Signage Players and specialized software that manage content playback. Alongside general information like news or video tutorials, digital signage screens also display ads, which can take up 100% of the screen time.

Our clients turned to 3DiVi Face SDK, which offered customization and smooth integration with their existing apps.

3DiVi Face SDK supports both 32-bit and 64-bit Android systems, ensuring easy implementation even on older devices. Integration was done using the Java API in both cases.

To optimize computational resources, we used the VideoWorker module, which featured:

• A lightweight face tracking algorithm.

• An image quality assessment tool that triggered data capture only when the face image was clear enough for template extraction.

The challenges we encountered were mainly due to running multiple computer vision models on devices with lower processing power, particularly older Signage Players. We addressed these issues by:

• Optimizing the configuration of 3DiVi Face SDK modules.

• Reviewing and refining the client's code to ensure efficient application of 3DiVi Face SDK modules.

The setup consisted of two key components:

1️⃣ Local Installation:

• Screen displaying the ads
• Camera positioned to capture the audience
• Ad player app that used 3DiVi Face SDK to analyze the video stream from the camera.

2️⃣ Server Integration:

• The app sends collected data to a server that builds data analysis pipelines to assess ad effectiveness.

By capturing rich audience data, the clients were able to optimize screen placement and tailor content more accurately to target audiences which led to a revenue boost of up to 22%, depending on the location and advertising content.

The integration of 3DiVi Face SDK brought invaluable insights and significantly improved the advertising effectiveness. The result? Better audience engagement and a notable revenue spike.

Want to boost your Digital Signage profits with AI-based Audience Analytics? Reach out for a consultation, and get a solution tailored to your needs.Integration of 3DiVi Face SDK and testing at multiple locations can be completed in just 1–1.5 months.