https://github.com/martinkersner/acceleration-of-2d-convolution-using-integral-image

Acceleration of 2D convolution using integral image
https://github.com/martinkersner/acceleration-of-2d-convolution-using-integral-image

Last synced: 15 days ago
JSON representation

Acceleration of 2D convolution using integral image

• Host: GitHub
• URL: https://github.com/martinkersner/acceleration-of-2d-convolution-using-integral-image
• Owner: martinkersner
• Created: 2013-04-24T12:21:33.000Z (over 11 years ago)
• Default Branch: master
• Last Pushed: 2015-10-13T10:22:42.000Z (over 9 years ago)
• Last Synced: 2023-06-12T23:10:19.890Z (over 1 year ago)
• Language: Python
• Homepage: https://dl.dropboxusercontent.com/u/13642345/Acceleration-of-2D-Convolution-Using-Integral-Image.pdf
• Size: 141 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## Acceleration of 2D Convolution Using Integral Image

Martin Kersner, 

[Full text](https://dl.dropboxusercontent.com/u/13642345/Acceleration-of-2D-Convolution-Using-Integral-Image.pdf)

Computation time of box blur filter using convolution depends on the width of kernel. Wider kernels causes longer computation time. However, there is a method of calculating blurred image in constant time called integral image. Due to the constant reading of values, it was reached almost 800 times acceleration with this method. High order filters can be computed by repeated box filtering. This paper also deals with implementation details of acceleration of 2D convolution using

integral image.