While you and I may find it easy to recognize faces amongst the many in public or the locations we come across on the Web, but computers have a tough time. There has been some work done to improve facial recognition, Google has recently shown interest in teaching computers to recognize landmarks. With this researchers at Google put out a few pictures of an unnamed, untagged picture of a landmark and the system returned it with its locations and name. The system matched the photo with more than 40 million images stored at Picasa and Panoramio along with those found on Google Image Search. The system is pretty intelligent, given that it recognizes more than 50,000 landmarks with 80% accuracy, but to put it out to test it has to do even better primarily because the number of locations available, a 20% shortcoming is a lot more than one would want.
-
-
Recent Entries...
- Mobile Payment Models An Overview–Infographic
- Project Management Made Easy With Teamwork Live!
- Collabor8Online Is About Taking Project Management Online
- Massive Online Collaboration: ReCaptcha The Captcha
- Startup: Grouperly Lets You Socialize With People Sharing Similar Interests
- Startup: Shphoonkle Is Your Free Legal Marketplace
- Startup: RTBot Gives You Access To All Real Time Information
- Infographic: What Happened To Ringtones?
- FileFriend Is Your Online File Manager
- How to Create Custom Shortcuts on iPhone Without Jailbreak
-
Random Entries...
-
Popular Tags
acquisition ads Android app apple Apps Ban China Chrome dead facebook funding Gmail google how to iOS iPad iphone iphone 4 iphone 5 ipod touch jailbreak microsoft mobile music myspace Nokia pakistan privacy Reports Rumors search social media social network startup startups tips tool tools Twitter verizon video web yahoo youtube



1 Trackback or Pingback for this entry:
[...] Google trying to make computers intelligent, helping them recognize landmarks Bewertung: 3.00 von 5 [...]