VisualSize, a Southern California based company, will be launching shortly with a service that will give accurate three-dimensional measurements of things inside of 2D photos.
To measure things accurately they need two photos of the same thing, but from different angles. The VisualSize algorithm automatically detects feature points in the two pictures and finds the matching pairs. It then uses the matching pairs to calculate coordinates of the two camera positions (x, y, z axes and origin in a 3D coordinate frame), and uses triangulation from the image to plane to the 3D coordinate frame to reconstruct the 3D scene. VisualSize can then measure length, angle, area and volume with a high degree of accuracy. The screen shot shows an example of raw output, but the service itself is pre-alpha.
This isn’t as sexy as, say, what Fotowoosh is doing by creating 3D images from a single 2D photo. But it does have commercial value. Home improvements, interior design, general contracting estimates, crime scene investigation, military planning, insurance claim assessments, etc. all need accurate measurements. VisualSize says they can create them with measurement errors of 0.1 to 5%.
The company says they will offer the service on their website. More importantly, there will be an API to allow others to build applications around the service.
The founders are Yuan-fang Wang (a UCSB professor) and Ping Liang (a UCR professor). They are self funded via liquidity events from previous startups.