Browse Prior Art Database

Fast and Effective Spike-Noise Removal Algorithm for Digital Images

IP.com Disclosure Number: IPCOM000123136D
Original Publication Date: 1998-Jun-01
Included in the Prior Art Database: 2005-Apr-04
Document File: 1 page(s) / 50K

Publishing Venue

IBM

Related People

Horiuchi, Y: AUTHOR

Abstract

Disclosed is an algorithm for removing Spike-Noise on a digital image more efficiently, and with less computation time. This algorithm automatically controls the pixels or sub-image areas that are applied to noise removal operation. To remove noise from only the target area maintains the quality of other areas.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 61% of the total text.

Fast and Effective Spike-Noise Removal Algorithm for Digital Images

   Disclosed is an algorithm for removing Spike-Noise on
a digital image more efficiently, and with less computation time.
This algorithm automatically controls the pixels or sub-image areas
that are applied to noise removal operation.  To remove noise from
only the target area maintains the quality of other areas.

   This algorithm is based on the Median-Filtering with
post-processing, a Pixel-Selection operation, that sorts out the
effective result from all filtered pixels.  To choose the effective
result, a predefined color value and an allowable distance is used;
If the distance between color value of a filtered pixel and the
predefined color value is shorter than the predefined distance, the
filtered pixel is used in the result image.  If not, it is discarded.

   Processing the Median-Filtering and the Pixel-Selection
simultaneously makes this algorithm fast.  The median operator can be
divided into several small stages; in each stage, a group of values
are chosen, and both maximum and minimum values are discarded from
the group.  The maximum and minimum values are computed by swap
operations.  Assume that the window size of the Median-Filter is 3x3,
and the color value is represented in RGB color space.  Then the
median operator will be applied for 9 values, and for each red, green
and blue component.  In this case, the operation is divided into 4
stages, and the number of values in the g...