CSCE 441: Computer Graphics

Image Filtering

Jinxiang Chai
 Outline
Image Processing
- Gaussian filtering
- Median filtering
- Bilateral filtering
 Filtering
• In signal processing, a filter is a process
that removes from a signal some
unwanted component or feature
1D Signal Filtering
2D Image Filtering
2D Image Filtering
 Image Filtering
Image filtering: change range of image
g(x) = h(f(x))
f f
h
x x

Image warping: change domain of image

g(x) = f(h(x))
f f
h
x x
 Image Filtering
Image filtering: change range of image
g(x) = h(f(x))

f g
h

Image warping: change domain of image

g(x) = f(h(x))
f g
h
Filtering in Spatial Domain

Filtered image Input image Filter function

Gaussian Filtering in Spatial
Domain

Discrete approximation to
Gaussian function with σ =1.0
Filtering in Spatial Domain

Discrete approximation to
Gaussian function with σ =1.0
Filtering in Spatial Domain

Discrete approximation to
Gaussian function with σ =1.0
Filtering in Spatial Domain

Discrete approximation to
Gaussian function with σ =1.0
 Filtering in Spatial Domain

Discrete approximation to
Gaussian function with σ =1.0

Filtered_I45 = 
pixelswindown
X
 Filtering

input Gaussian filter

 Median Filter
• For each neighbor in image, sliding the
window
• Sort pixel values
• Set the center pixel to the median
 Median Filter

input Gaussian filter Median filter

Median Filter Examples

input Median 7X7

Median Filter Examples

Median 3X3 Median 11X11

Median Filter Examples

Straight edges
kept

Median 3X3 Median 11X11 Sharp features

lost
 Median Filter Properties
Can remove outliers (peppers and salts)

Window size controls size of structure

Preserve some details but sharp corners

and edges might get lost
Comparison of Mean, Gaussian,
and Median

original Mean with 6 pixels

Comparison of Mean, Gaussian,
and Median

original Gaussian with 6 pixels

Comparison of Mean, Gaussian,
and Median

original Median with 6 pixels

 Common Problems
Mean: blurs image, removes simple noise, no
details are preserved
 Common Problems
Mean: blurs image, removes simple noise, no
details are preserved
Gaussian: blurs image, preserves details only for
small σ.
 Common Problems
Mean: blurs image, removes simple noise, no
details are preserved
Gaussian: blurs image, preserves details only for
small σ.
Median: preserves some details, good at removing
strong noise
 Common Problems
Mean: blurs image, removes simple noise, no
details are preserved
Gaussian: blurs image, preserves details only for
small σ.
Median: preserves some details, good at removing
strong noise

Can we find a filter that not only smooths regions

but preserves edges?
 Common Problems
Mean: blurs image, removes simple noise, no
details are preserved
Gaussian: blurs image, preserves details only for
small σ.
Median: preserves some details, good at removing
strong noise

Can we find a filter that not only smooths regions

but preserves edges?
- yes, bilateral filter
 Outline
Image Processing
- Gaussian filtering
- Median filtering
- Bilateral filtering
 What Is Bilateral Filter?
Bilateral
- Affecting or undertaken by two sides equally

Property:
- Convolution filter
- Smooth image but preserve edges
- Operates in the domain and the range of image
Bilateral Filter Example

Gaussian filter Bilateral filter

Bilateral Filter Example

Gaussian filter Bilateral filter

 1D Graphical Example
Center
Sample u
I(p)

Neighborhood
p  neighbor (u )

It is clear that in weighting this neighborhood,

we would like to preserve the step
 The Weights

I(p)

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 Filtered Values

I(p)

Filtered value

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 Edges Are Smoothed

I(p)

Filtered value

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 What Causes the Problem?

I(p)

Filtered value

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 What Causes the Problem?
Same weights for these
I(p) two pixels!!

Filtered value

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 The Weights

I(p)
( I (u )  I ( p)) 2
Ws ( p)  exp(  )
2 s
2

(u  p) 2
Wc ( p)  exp(  )
2 c
2

p
 Bilateral Filtering
Denoise Feature
preserving
Bilateral filter
u p
2
I u  I  p 
 


pN  u 
e 2 c2
e 2 s2
I ( p)
Normalization
I ' u   2
I u  I  p 
u p
 


pN  u 
e 2 c2
e 2 s2
 Kernel Properties

I (u ) 
 W ( p ) *W ( p ) * I ( p )
p c s

 W ( p ) *W ( p )
p c s

• Per each sample, we can define a ‘Kernel’ that

averages its neighborhood
• This kernel changes from sample to sample!
• The sum of the kernel entries is 1 due to the
normalization,
• The center entry in the kernel is the largest,
• Subject to the above, the kernel can take any
form (as opposed to filters which are
monotonically decreasing).
 Filter Parameters
I (u ) 
 pN ( u )
Wc ( p) *Ws ( p) * I ( p)
 pN ( u )
Wc ( p) *Ws ( p)

As proposed by Tomasi and Manduchi, the filter is

controlled by 3 parameters:
N(u) – The neighbor size of the filter support,
c – The variance of the spatial distances,
s – The variance of the value distances,

The filter can be applied for several iterations in order to

further strengthen its edge-preserving smoothing
 Bilateral Filter

input
 Bilateral Filter

input
 Bilateral Filter

input

Wc
 Bilateral Filter

input

Wc Ws
 Bilateral Filter

input

Wc Ws Ws*Wc
 Bilateral Filter

input Output

Wc Ws Ws*Wc
Bilateral Filter Results

Original
Bilateral Filter Results

σc = 3, σs = 3
Bilateral Filter Results

σc = 6, σs = 3
Bilateral Filter Results

σc = 12, σs = 3
Bilateral Filter Results

σc = 12, σs = 6
Bilateral Filter Results

σc = 15, σs = 8
 Additional Comments
The bilateral filter is a powerful filter:
- Can work with any reasonable distances
function Ws and Wc,
- Easily extended to higher dimension signals,
e.g. Images, video, mesh, animation data etc.
- Easily extended to vectored-signals, e.g. Color
images, etc.

Bilateral Mesh Denoising

[Fleishman et al, siggraph 03]