ZunaVision is an up and coming technology that enables advertisers to dynamically display ads inside the 3D space of online videos. Zunavision was developed by 2 students currently attending Stanford University, Ashutosh Saxena and Siddharth Batra. I had the chance to sit down with both of them to get some insight on ZunaVision and their ideas behind developing it.
[Manoj]: What is your educational background?
[Ashutosh & Siddharth]: Ashutosh Saxena received his undergraduate degree from Indian Institute of Technology Kanpur (IIT Kanpur, India) in 2004, and his Masters degree from Stanford University in 2006. Currently, he is finishing his PhD at Stanford University in the area of Machine Learning and Computer Vision.
Siddharth Batra is finishing his Masters degree in Computer Science (with a distinction in computer vision research) from Stanford University, and received his undergraduate degree in 2007 from JIIT (Jaypee Institute of Information Technology) in India. His area of expertise include Computer Vision and Computer Graphics.
[Manoj]: Where did you get the inspiration to create ZunaVision?
[Ashutosh & Siddharth]: The online video ecosystem is devoid of an efficient advertising & monetization mechanism and existing approaches which emulate television standards have not done very well. The solution to this problem and the idea of using our computer science expertise towards changing the way the world thinks about video advertising is our inspiration and the aim of our company.
[Manoj]: Can you give us a brief overview of how it works?
[Ashutosh & Siddharth]: Our company’s product enables advertisers to dynamically display ads inside the 3D space of online videos. Content owners and producers take seconds to make their video ad-friendly via either our website or one of our partner video distribution websites.
The resultant ad-unit is an uninterrupted and natural viewing experience for viewers; an unskippable form of advertising for online marketers; and increased monetization for video producers and sites
Behind this lies a technology that we have developed in the Stanford AI lab over the last few years. It uses machine learning techniques that learns the 3D structure of the videos.
Previously, we had developed Make3D (http://make3d.stanford.edu/publications.html), that allows a computer to estimate 3D structure from a single image. We have extended this to videos, where the algorithm has to infer how the rectangle (where the ad is to be placed) moves in 3D.
[Manoj]: When is the commercial version being launched?
[Ashutosh & Siddharth]: In the first week of February.
[Manoj]: How much interest have you had for the commercial version?
[Ashutosh & Siddharth]: We are getting significant global interest across all verticals in the online video ecosystem including content producers, ad-agencies and video distribution websites. Updates about the commercial release will be posted at http://www.zunavision.com.
[Manoj]: Have you determined your monetization strategy? How will you determine what you charge for your service and how much advertisers will pay?
[Ashutosh & Siddharth]: We have a core set of revenue models for different types of clients, and the parameters of these models vary for each of them. We believe this flexible yet apt monetization strategy will be one of the key factors in the success of our company.
Here is a sample of ZunaVision in action: