Nonlinear Filtering for Image Analysis and Enhancement Gonzalo R' Arce, Jose L' Paredes and John Mul

1
Nonlinear Filtering for Image Analysis and
Enhancement Gonzalo R. Arce, Jose L. Paredes
and John Mullan

• Kim Mose
• Electronic Engineering
• Sejong UNIV.

2
1. Introduction

• Image processing is being used in many areas such
  as industrial or scientific applications,
  military, internet, digital camera and etc.
• Weighted Median, nonlinear method, is very
  effective to preserve edges and details of image

Image Processing
Enhancement
Analysis
Nonlinear Method
Linear Method
3
2. Weighted Median Smoothers and Filters (1)
4
2. Weighted Median Smoothers and Filters (2)

• Recursive median smoother is more cost effective
  than general medians, but not robust
• Median smoother is very useful in Laplacian noise
  rather than Gaussian

5
2. Weighted Median Smoothers and Filters (3)
6
2. Weighted Median Smoothers and Filters (4)

• Real-valued WM smoother
• Calculate the threshold
• Sort the samples in the observation vector x(n)
• Sum the weights corresponding to the sorted
  samples, beginning with the maximum sample and
  continuing down in order
• The output is the sample whose weight causes the
  sum to become greater than or equal to To

7
2. Weighted Median Smoothers and Filters (5)
8
3. Image Noise Cleaning

• 5x5 CWM apply to impulse noised image
• (a) is original image
• (b) is noised image
• (c) is Wc 15
• (d) is Wc 5

9
4. Conclusion

• Its apparent that nonlinear filter also provides
  good image enhancement
• WM is very powerful to preserve the edges and
  details of image
• WM is also capable of applying to variety of
  image processing areas