Visor: Object recognition framework for augmented and virtual reality devices

Visor aimed to develop an object recognition framework adapted to glasses-type augmented reality devices, given the boom these devices were experiencing, as well as a virtual environment modeler for virtual reality devices.

The project's objectives were:

• To produce a generic set of algorithms for recognizing objects from images.
• To implement the framework on a specific device, such as smart glasses, adapting the algorithms to run in real time on devices with limited computing capabilities.
• To develop facial attribute detection.
• To develop image-type marker detection.
• To model existing environments in tools such as CAD3D or Revit and transfer them to Unity for display on virtual reality devices such as Oculus Rift.

The core areas of Visor were:

• Object recognition: algorithms for robust extraction of information from images, taking into account factors such as changes in lighting, shadows, pose, scaling, or rotation.
• Implementation on portable devices: adaptation of the algorithms to run in real time on augmented reality glasses (Google Glass) and other Android devices (phones, tablets, smartwatches).
• Virtual environment modeling: conversion of CAD3D or Revit models to Unity for display on virtual reality devices such as Oculus Rift, reducing the number of polygons without losing model detail.

CTIC developed Visor with its own funds and with the support of the Asturias Program for Technology Centers.