Humans have long-dreamed of flying around the world and exploring never before seen locations or remote historical sites restricted to visitors. Of course, the reality is that the ability to do this is limited by gravity, ability, time, and money. But imagine if there’s a way to overcome these limitations and have similar experiences either without or before actually doing them?
As it turns out, in recent years, this dream has become possible through virtual tours and street views. However, creating these tours requires a lot of data, processing time, and expertise. What’s more, how can these environments be built if there are only a handful of photographs or images of the location one wants to explore?
Intelligence Advanced Research Projects Activity (IARPA) Program Manager, Ashwini Deshpande, believes the solution to these problems is within reach with the Walk-through Rendering from Images of Varying Altitude (WRIVA) program. WRIVA’s goal is to develop software algorithms that can build photorealistic environments using a few images taken from a ground-viewpoint, along with images from traffic cameras and satellites.
"With the recent advances in rendering and view synthesis using machine learning, the prediction of missing viewpoints has become technically possible and would allow us to create seamless, photorealistic site models in areas that are hard to access,” Ms. Deshpande said.
If successful, WRIVA will be able to take imagery from different sources, predict novel views to fill in the gaps and render a high-fidelity ground-level immersive environment to familiarize users with a new area. The intent is to help users navigate to determine the best path to deliver humanitarian aid or for any other use where one needs help visualizing a destination and how to get there before actually attempting to travel to and arrive at the location.
For the Intelligence Community (IC), military, and law enforcement, WRIVA’s potential operational impact is significant. "Situations where you need to deliver support quickly but also need to understand the area you are going to is where WRIVA could be most impactful.” Ms. Deshpande said.
WRIVA will be able to help plan missions and other activities with highly-accurate virtual environments that can be produced quickly using images taken by non-experts. This will provide officials virtual “ground truth” and help them prepare to operate in-person with more confidence.
“The ability of WRIVA to help users visually see and plan a mission or activity, even with limited access to imagery, will be a game-changer for the IC, military, and law enforcement,” Ms. Deshpande said. “And while it won’t be better than reality, it might mean the difference between mission failure or success.”
