http:^^www.cs.cornell.edu^info^people^scytacki^research.html
来自「This data set contains WWW-pages collect」· HTML 代码 · 共 39 行
HTML
39 行
MIME-Version: 1.0
Server: CERN/3.0
Date: Tuesday, 17-Dec-96 00:00:26 GMT
Content-Type: text/html
Content-Length: 1604
Last-Modified: Saturday, 16-Dec-95 07:01:22 GMT
<html><head><title>Research</title></head><body background="images/tan2.jpg"><h1>Stuff</h1>Currently I am working on an implementation of a image database.The implementation that we are trying to do is one written in First in Rivland latter in Visual C or Visual Basic. The idea of this database is to use as many images as we can handle fromthe web, and run a program on each one which generates an electronic discription of the image. Then a second program will take this database and a sample image and generate a list of the top n images with similardescriptions.<br> There are several similar databases out thereright now. Links to those will come latter. Most of these databases use color histograms to describe their images. A color histogram is a listof colors and the amount of each color in the image.Color histograms are good for certain things, but a little example with demonstrate where they fail.<br>Suppose you have a picture of a sunset. (picture is coming). If this wascompaired to a randomly generated image that resemble the snow you sometimessee on TV(but in color), the result could be a perfect match. It is obviousthat it is not a perfect match, but each images has exactly the same amountof each color in it.<br>The goal then of this database is examine other methods of compairing imagesthat better represent a humans perception of similarity.<br>Currently we have a comparision which incorporates positional data as well asthe color data. Hopefully I have a better description of this later. As longas it can be released. </body></html>⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?