Digital Signage Gets Physical - Part 4: Face Detection and Audience Analytics

This blog series has explored the various ways in which digital signage can interact with it's environment. In this post we will focus on the audience and how it's possible to adapt signage content in real-time based on viewer attributes. It's important to recognize that the data collected in this process is completely anonymous and can in no way be used to track individual viewers.

What is face detection?

By using what's called 'computer vision', it's possible to algorithmically determine the number of faces in a video frame along with approximations for age, gender, and even mood. This data can then be used in real-time to adapt signage to better match the make-up of the current audience. For example, if the audience is primarily male, then adjust the playlist to include more male oriented content. By targeting the right demographic you are automatically improving the value of your content.

What is audience analytics?

Audience analytics is taking the audience data and massaging it to learn various 'big picture' statistics and trends. Revel Digital has a suite of visualization tools we refer to as AdHawk which can provide these valuable insights.

AdHawk provides audience analytics

With AdHawk it's possible to determine, at a glance, various metrics such as total viewership, busy times of the day, age and gender percentages, dwell time, and more. You can even compare those metrics with previous periods for A/B testing. Combined with proof-of-play reports, this data is invaluable to advertisers.

Getting Started

Most any Android device with a camera should work well. The Revel Digital player app for Android is all you need to install. For those of you new to Revel Digital, please see our website for more information.

Once your player is up and running you'll need to open the player settings by pressing Shift-F12 on an attached keyboard. Then click the Camera Mode option and select Audience Analytics.

Enable audience analytics in the player settings

That's it! your player is now collecting audience data in real-time.

Demo

Our demo will utilize the real-time face detection to tune a playlist. The playlist will consist two sets of images, one set will target male viewers, the other set will target female viewers. When the majority of viewers are male the playlist will show more male oriented images, and vice-versa for female viewers.

Step 1: Create a Slideshow Playlist

Step 2: Add at least 2-3 different images.

For one of your images, click the Schedule tab and add a When condition for '% Male Viewers'. Enter a value of 50. This means the media will play only when at least 50% of the viewers are male.

Do the same thing for another playlist item, but this time use the '% Female Viewers' condition.

Step 3: Schedule the playlist to your player device and update.

You're done! Now the media will only appear when the associated audience condition has been met. This same technique can be used for any playlist item, and anywhere smart scheduling is available.

Another popular use case for face detection is for attract screens. A kiosk, for example, could play an attract loop until a viewer has been detected.

AdHawk Analytics

Audience data will be available in your AdHawk tools as well, but only for devices on the Professional+ or Enterprise+ tiers. You can adjust your tier but editing a device and clicking the Change Service Level button.

Audience analytics is available to the Plus service levels

Conclusion

Technology is quickly becoming more intelligent and digital signage is no exception. It's important for digital signage owners to get their content in front of the intended audience and this technology is great means to that end.