We were looking to provide a more personalized ad experience to TV streams, like how currently internet ads work. In order to do that we had to remove the non-relevant ad and display the personalized ad. In addition provide access to retrieve information about the viewed advertisement.

The server receives live feed and analyzed the video. After the server analyzed the video in real time it sends messages to the client to switch the ad or provide a more through experience by letting the user click on a web action. With this platform we are able to provide the internet advertisement experience to TV and video content.

Challenges I ran into

Issues setting up platforms for work, spent a lot of hours installing open CV and elastic beanstalk. issues parsing HLS feed, and analyzing the video feeds.

Accomplishments that I'm proud of

Our progress on the project, we are able to analyze a video feed and get accurate response to commercials. As well as a small team of 2 people working on this project was a challenge.

What I learned

Different open CV algorithms, using elastic beanstalk, setting up open cv for node.

What's next for Commercial Personalizer

Being able to fully analyze commercial logos so we can provide the relevant info and action clicks. Improve our accuracy on the video stream, by testing different algorithms. Add support for more TV stations.

Share this project: