Advertising is a huge market and a major part of our economy. Although technology has made advertising online— cheaper, more efficient, and more relevant—little progress has been made with offline ads.
Offline ads like billboards, posters, and signs have the potential to reach many, many people— but when was the last time any of these were relevant to you?
ARISE was inspired by this problem— after all, why shouldn't offline ads be personalized to each viewer, just like online ads? We set out to leverage augmented reality (AR) to accomplish this end. We wanted to develop a mobile app that would enable marketers and community leaders to leverage the same offline ads in an interactive and virtual manner that is relevant to each user.
What it does
ARISE redefines the offline advertising user experience by generating a personalized advertisement relevant to you.
A user simply scans an AR code (e.g. on a billboard, or poster) and is automatically presented with relevant offers and recommendations based on their online footprint.
ARISE aims to help both the user and the marketer. Users are spared from viewing irrelevant ads and control their experience - users are incentivized by marketers to scan an AR code (e.g. via deals). Marketers obtain higher quality leads at a more economical cost by enabling multiple marketers to access the same placement.
ARISE advertisements have a myriad of use cases from product marketing to spreading awareness for community or environmental initiatives.
How we built it
The front-end is a native Android app written in Java, and heavily reliant on Google's ARCore. We are able to take an "AR code" and overlay a targeted advertisement using AR technologies. We use an
AugmentedImage to overlay our own advertisement image along with descriptive Text (check out our video)!
The back-end is a Flask app written in Python, and heavily reliant on Selenium. We wanted to provide real personalized ads, so we leveraged Facebook and Amazon's targeting to illustrate proof of concept. Essentially, we web-scrape targeted ads from these platforms and display them when a user scans an AR code.
Challenges we ran into
- Getting ARCore to render 2D images and text on a specific point (in a way that looks right); when we were able to render 2D images they were "mirrored" and we had to account for that
- Web scraping from Amazon was not possible due to two-factor authorization and other security measures.
- Instagram did not show advertisements on the desktop version.
Accomplishments that we're proud of
- It works!
- We were able to create a functioning AR app despite having no prior experience with AR
What we learned
What's next for ARISE
- Building our own advertisement personalization algorithm so that we do not have to rely on third-parties like Facebook, Amazon, or Google.
- Implementing an advertisement rating system to fine-tune personalizations.
- Developing a reward and point incentivization system for the user.
- Possibly use a decentralized method for building personalized advertisements.