eople “are taking more photos than ever. We don’t wait to capture just that right moment; we can capture EVERY moment. Our photo histories are larger and more complex. We share these photos in new ways, by email, by cell phones and on the Internet.
However, with these growing photo sets comes a growing mess. We’ve moved our photos from stacks of shoeboxes to our hard drives, where we have a bigger, more complicated mess to sort through to find the photos we want to send.
These photos are your stories, your moments and your experiences.The Riya vision is to make it easier for people to tell their stories with their pictures. Our photo search helps you do this.
Riya is more than photo search. Our goal is to help you find every photo of yourself on the web. We want to help you recover every moment, every place you’ve been and all of the people you’ve met along the way. We want to give you the tools to discover your future, every place you want to go and meet new friends. We will be successful when we can find every digital photo in the world. [...]
Computer vision can be described as the study of methods which can be used for allowing computers to ”understand“ images, or multidimensional data in general.
Over 30 years of research has gone into the study of Computer Vision, but it is still a field largely in its infancy. The largest contributing factor to this slow advancement is that the process that extracts 3-D information from 2-D images is much more complex than the flipside science, Computer Graphics. A computer has an easier time translating 3-D information into 2-D imaging.
In addition to the overall complexity, the lack of consumer application of the process slows down progress of development...until now.
Face recognition has been heavily studied within the field of Computer Vision over the past decade. The difficulties of face recognition lie in the inherent variation of facial characteristics and the environment of image acquisition. The following summarizes the major variations in the images of a single individual:
- 3D poses, image scaling, and view point
- Changes in the actual faces: closing of the eyes, wearing glasses,facial expression, facial hair, aging, make-up, and perspiration
- Partial occlusion
- Light source strength, direction, and distribution
Most of the existing face recognition systems to date usually impose restrictions to avoid those difficulties on the imaging situation, such as well-controlled environments or near-frontal, well-aligned images.
However, how many digital photos do you have that are shot straight-on in well-controlled environments? Exactly. That is why our team has been working so hard at getting the balance just right in achieving highest level of accuracy with the highest level of recognition possible.”
[Riya]