http://www.nytimes.com/2012/11/20/science/for-web-images-creating-new-technology-to-seek-and-find.html?ref=technology
On November 19, 2012 John Markoff wrote an artile on ways you can better find the image you are looking for. Now, you find an image through search engines and the only images that pop up is what matches what you typed in the search box with the picture caption. So many times it is hard to find the exact image you are searching for. Therefore scientists will create a new generation of visual search technologies before the interent goes "dark" as scientist Fei-Fei Li said. With the help of many scientists, Dr. Li created the world´s largest visual database with over 14 million objects. Humans wrote the lables but now the machines can recognize and make good assumptions as to what object it is. It has been tested, known as Imagenet, and worked very well. This is a big site not only for Google but also many scientists. Dr. Li´s goal is to build a vision system the way a human´s brain works. The interent is about to explode with all the imagens and videos we are uploading in it. Dr. Li found out that the Mechanical Turk was a perfect way to gather her database. Every year Imagenet provides images for 20,000 to 30,000 people. A normal worker can indentify 250 images in 5 minutes. The futhur up in the caterogy you go, the less error you will find. They are looking to getting better and better as time goes.
The purpose of this article to show the reader another option in which you can search an image that you want and it will pop up. Markoff gives a good explaination as to what Imagenet does and how much better or accurate the search is compared to other search engines. Markoff provides us with many well said quotes on how it works and why it is better than anything else out there right now.
No comments:
Post a Comment