Next-Generation Web Searches for Visual Content

  • Authors:
  • Michael S. Lew

  • Affiliations:
  • -

  • Venue:
  • Computer
  • Year:
  • 2000

Quantified Score

Hi-index 4.10

Visualization

Abstract

Although major search engines facilitate finding text on the Web, they typically have few or no capabilities for finding visual media. Many Web users--such as magazine editors or professional Web site designers--need to find images by using a few global features. With hundreds of millions of sites to search through, and 73 percent of the Web devoted to images, finding the exact image needed can be a daunting task.The author and his colleagues developed a prototype system called ImageScape to find visual media over intranets and the Web. The system integrates technologies such as vector-quantization-based compression of the image database and k-d trees for fast searching over high-dimensional spaces. The author claims that ImageScape is the only Web search engine that allows sketch- and icon-based queries.Commercial and academic institutions are working on new paradigms for visual searches including searching by icons, sketches, and similar images. These paradigms have the potential to bring the majority of Web information to anyone with a browser. The author projects that future research will focus on the fusion of visual learning techniques such as neural net-works and decision trees, combining them toward improving visual-concept detection accuracy.