A Case Study on Interactive Product Displays at Events

In today’s crowded brand activation landscape, every brand is looking for a way to stand out, engage, and educate all at once.

At ANTIGRVTY, we help brands transform ordinary booths into smart, interactive experiences powered by computer vision. One of our favorite recent projects shows how a simple setup. Two screens, a few sensors, and smart AI can completely change how people interact with products.

🧠 The Challenge

At events, brands like CeraVe often want consumers to learn about their products, not just see them. But physical displays are usually static, visitors walk by, maybe pick up a product, and move on without remembering the details.

🤖 The Idea: What if the product could introduce itself?

We created an interactive display using two synchronized TVs:

  • Lower screen: A TV surface equipped with a camera and computer vision.
  • Upper screen: A main visual display that dynamically changes content.

When a visitor places a real product (for example, CeraVe Blemish Control Gel) on the lower screen, our computer vision system instantly recognizes it.no barcode scan, no NFC tag and triggers a customized visual on both screens.

🧪 Behind the Scenes: Training the Vision System

Before the live activation, our team went through an intense training phase.
We didn’t just use generic AI, we trained our own computer vision modelspecifically for the CeraVe product line.

Each product — whether it’s Blemish Control Gel, Retinol Serum, or Foaming Cleanser — was photographed from multiple angles, under various lighting conditions.

On average, we trained the AI using around 300 photos per product.

Testing our computer vision system recognizing CeraVe products in real time.

This allowed the system to identify each SKU with near-perfect accuracy, even when the product was rotated, slightly tilted, or partially shadowed.

We tested it repeatedly — placing products in random positions, different distances, and even under different lighting setups — to ensure smooth real-world performance at the booth.

By the time it went live, the system was stable, instant, and felt alive.