EdgeDetector 1.0 & 1.1

EdgeDetector is a program that detects edges from a gray scale image in GIF format. Version 1.0 was made available in April 1999, while Version 1.1 was made available in June 1999. All versions have been tested on Windows 98/NT and Solaris.

How to use EdgeDetector:

  • Download EdgeDetector 1.0 or 1.1.
  • Use the original head image, or one of the smoothed images or another image that meets the requirements for the input image specified at the bottom of this page.
  • Type: ED InFileName OutFileName LL ES. Here:
    • ED is "EdgeDetector", in case of EdgeDetector 1.0; or "EdgeDetector11" in case of EdgeDetector 1.1.
    • InFileName is the name, including the path, of the input image file name. It must end with the extension of ".gif".
    • OutFileName is the name, including the path, of the result image file name that contains the detected edges. The output image will be GIF format. Thus the file name should end with the extension of ".gif". If not, such an extension will be added to the specified file name.
    • LL denotes the lower limit of the edge contrast. It can be any number no smaller than zero and greater than 100.0
    • ES denotes the edge strength. It can be any number no smaller than zero and greater than 100.0
  • Please email me (wwzhou at yahoo dot com) your comments/suggestions. If you would like to be notified about future releases of EdgeDetector, please say so.

The effect of the last two parameters, LL and ES are demonstrated with examples. Both of them are used in a relative sense. In other words, edges with the same contrast and strength may not be detected from two different images with the same values for LL and ES, depending on the number of edge pixels present in the images. With the same values for LL and ES, EdgeDetector 1.0 and EdgeDetector 1.1 usually detect different numbers of edge pixels from the same image.

Generally speaking, the greater the values of LL and ES are set, the more pixels will be highlighted in the output image as edges. If either LL or ES is set to zero, no edge in the image will be considered strong enough to be highlighted. If both LL and ES are set to 100, all pixels that compose an edge, no matter how weak it is, will be highlighted in the output image.

If LL is set to 100, EdgeDetector can be used to obtain the results shown in the examples by appropriately setting values for ES and smoothing the original images.

Requirements for the input image:

  • The input image should be in GIF format.
  • Either dimension of the input image should not be greater than 256.
  • The input image should be a gray scale image. If it contains three bands, only the red band will be used. In other words, edges visible only in the green and blue bands will not be detected.

Edge Detection Home Page
Results from the head image || Results from the Lena image
Download EdgeDetector || Download Synthetic Test Images


Other Projects