Tag your photo album and retrieve instantly

Riya Photo Search, a Silicon Valley startup which has gone the farthest in developing digital search technology for images, on Thursday said it has received a substantial $15 million of second round of funding. This will be utilised to strengthen its R&D centre in Bangalore.

Munjal Shah, CEO and co-founder, said, "R&D for the company will be headquartered in Bangalore." The company received a $4 million initial funding in venture capital from LeapFrog and BlueRun Ventures. The second round has come from Bay Partners.

Riya, which works on facial recognition technology, sorts and tags customer's snapshot. The technology allows photos to be automatically tagged when it recognises whose photos they are. The service will be launched next month. This will help customers who want to sort their huge personal albums.

The customer first uploads his album to the Riya website which "recognises" pictures and attaches tags to the. The pictures are copied and sent back. When more pictures are uploaded they are tagged according the previous tags already available. A customer can thereafter, by going to the website, retrieve say all his wife's pictures previously uploaded by typing the name. It also recognises texts attached to photos. This is different from Google technology whose image search uses the text typed and not facial recognition technology.

Facial recognition technology software is based on the ability to first recognise faces and then measuring the various features of each face. In the case of Riya, it can also recognise the texts in the photos. The technology is mainly useful for families having albums to sort out.

Riya will first launch its services under which customers will have to log in to Riya's website and then upload their photo albums. The software will index the photos. "At first, the customer will have to train the software but once trained it will pick up the pictures from the album," said Shah.

The "train the face recognition" software requires a downloadable Windows client which identifies and tags individual photos. As the software learns who is who in the picture, it autotags the photograph itself, quickly scanning the rest of the photos and identifying each person it recognises.

Riya also uses text recognition to read street and other texts in photos. Photos uploaded to the free service for tagging can be made private or users can choose to make images viewable. Munjal added, "Friends may search each others' photos and even add their own tags."

Shah said that as of now it does not plan to enter the internet like Google or Yahoo search but it will remain a future option.

Headquartered in Silicon Valley, the company works with a core set of engineers and presently employs 25, nine of whom are based at the Indian R&D centre in Bangalore. The company plans to double the number of engineers at the Bangalore facility.

The revenue model for the company will be ads at its website. This was also the company Google was supposedly considering for a possible purchase but it did not take place.