Chapter Overview: Point Operations!: Relating Gray Values To Brightness!

Chapter Overview: Point Operations !
Relating gray values to brightness!

Gray value quantization! Webers Law! Gamma characteristic! Adjusting brightness and contrast!
Gray-level histograms and histogram equalization! Point operations for combining images!

Averaging! Subtraction! The need for image registration!
Bernd Girod: EE368 Digital Image Processing!
Point Operations no. 1 !
Quantization: how many bits per pixel? !

8 bits
5 bits
4 bits

Contouring

3 bits

2 bits

1 bit

How many gray levels are required? !
Contouring is most visible for a ramp!
32 levels! 64 levels! 128 levels! 256 levels!
Digital images typically are quantized to 256 gray levels. !

Bernd Girod: EE368 Digital Image Processing! Point Operations no. 3 !
Brightness discrimination experiment !
Can you see the circle?!
I + I
125! 126! 128! 129! 130! 127!
Note: I is luminance, measured in cd m 2
Visibility threshold!
I I KWeber 12%
Weber fraction! Webers Law!

Contrast ratio without contouring !
Luminance ratio between two successive quantization levels at visibility threshold!
I max N 1 = (1 + KWeber ) I min

For! KWeber = 0.010.02 N = 256 Typical display contrast ratio!

I max = 13156 I min
Modern at panel display in dark room 1000:1! Cathode ray tube 100:1! Print on paper 10:1!
Suggests uniform perception in the log(I) domain (Fechners Law)!

Gamma characteristic !
Cathode ray tubes (CRTs) are nonlinear!

Luminance!
I!
I ~ U
= 2.0 . . . 2.3!
Voltage U!
Cameras contain -predistortion circuit!
U ~ I1
log vs. -predistortion !

U

U ~ log(I )
U ~ I1
Imax = 100 Imin
Similar enough for most practical applications!

Photographic lm!
Hurter & Drifeld curve (H&D curve)! for photographic negative! slope -!
Luminance!
shoulder!
2.0

density d

I = I 0 10 d = I 0 10 ( log E +d0 ) = I 0 10 d0 E
1.0

toe!
linear region!
log E
E is exposure

0
d0

measures lm contrast!
General purpose lms: = -0.7 . . . -1.0! High-contrast lms: = -1.5 . . . -10!
Lower speed lms tend to have higher absolute !
Brightness adjustment by intensity scaling !

Original image! Scaled image!
f [ x, y ]

a f [ x, y ]
Scaling in the -domain is equivalent to scaling in the linear luminance domain
I ~ ( a f [ x, y ]) = a ( f [ x, y ])
. . . same effect as adjusting camera exposure time. !
Contrast adjustment by changing !

Original image!
increased by 50%!
f [ x, y ]
a ( f [ x, y ])
with
= 1.5
. . . same effect as using a different photographic lm . . .!

Contrast adjustment by changing !
Scaled ramp 20!
Original ramp 0!
Scaling chosen to! approximately preserve! brightness of mid-gray!
Scaled ramp 0.50!

Gray level histograms !
Distribution of gray levels can be judged by measuring a histogram:!

For B-bit image, initialize 2B counters with 0 ! Loop over all pixels x,y! When encountering gray level f [x,y]=i, increment counter #

Normalized histogram may be thought of as an empirical probability distribution.! You can also use fewer, larger bins to trade off amplitude resolution against sample size.!
Example histogram !
#pixels!
Cameraman ! image! gray level!
Example histogram !
#pixels!
gray level!
Pout ! image!
Histogram equalization !
Idea: nd a non-linear transformation!
g=T(f)
!to be applied to each pixel of the input image f [x,y], such that a uniform distribution of gray levels in the entire range results for the output image g[x,y].
Analyse ideal, continuous case rst, assuming! 0 g 1 0 f 1
T(f) is strictly monotonically increasing, hence, there exists! 0 g 1 f = T 1 ( g )
Goal: pdf pg(g) = 1 over the range ! g 1 0

Histogram equalization for continuous case !
From basic probability theory!

df pg ( g ) = p f ( f ) dg f =T 1 ( g )
Consider the transformation function!

g = T ( f ) = p f ( ) d
0 f
0 f 1
Then . . . !
dg = pf ( f ) df
df 1 pg ( g ) = p f ( f ) = p f ( f ) =1 dg f =T 1 ( g ) p f ( f ) 1 f =T ( g )
0 g 1
Histogram equalization for discrete case !
Now, f only assumes discrete amplitude values f0 , f1 ,, fL1 ! !with empirical probabilities!
n0 P0 = n n1 P1 = n nL1 PL1 = n
L1
where n = nl
l =0
Discrete approximation of! g = T ( f ) = p f ( ) d

0
gk = T ( fk ) = Pi
i=0
for k = 0,1..., L 1
The resulting values gk are in the range [0,1] and need to be scaled and rounded appropriately. !
Histogram equalization example !
Original image Pout !
Pout! after histogram equalization !


Original image Pout ! . . . after histogram equalization !
#pixels!
gray level!
#pixels!
gray level!
Original image Cameraman !
Cameraman ! after histogram equalization !

Original image Cameraman ! . . . after histogram equalization !
#pixels!
gray level!
#pixels!
gray level!
Original image Moon !
Moon ! after histogram equalization !

Original image Moon ! . . . after histogram equalization !
#pixels!
gray level!
#pixels!
gray level!
Contrast-limited histogram equalization !
Output gray level!
Equalization with! original histogram!
#pixels!
Equalization with! clipped histogram!
Histogram Clipping!
Input gray level! gray level!
Input gray level!
Adaptive histogram equalization !
Apply histogram equalization based on a histogram obtained from a portion of the image!
Sliding window approach:

different histogram (and mapping) for every pixel!
Tiling approach:
subdivide into overlapping regions, mitigate blocking effect by smooth blending between neighboring tiles !
Must limit contrast expansion in at regions of the image, e.g. by clipping individual histogram values to a maximum (CLAHE - Contrast-limited adaptive histogram equalization)!
[Pizer, Amburn et al. 1987]!
Original!
Global histogram!
Tiling 8x8 histograms!
Tiling 32x32 histograms!
Original image Dental Xray !
Adaptive histogram equalization, 16x16 tiles !
Dental Xray after equalization of global histogram!
Original image !
Global histogram equalization !
Adaptive histogram ! equalization 8x8 tiles !
Point operations for combining images !

Image averaging for noise reduction! Combination of different exposures for high-dynamic range imaging! Image subtraction for change detection! Accurate alignment is always a requirement!
Image averaging for noise reduction!

1 image 2 images 4 images!
http://www.cambridgeincolour.com/tutorials/image-averaging-noise.htm!
High-dynamic range imaging !

16 exposures, one f-stop (2X) apart!
[Debevec, Malik, 1997] !

Combined image!
Image subtraction !

Find differences/changes between 2 mostly identical images! Example: digital subtraction angiography!
Contrast ! enhancement !
http://www.isi.uu.nl/Research/Gallery/DSA/!
Video background subtraction !

New Frame ! Background Frame ! New Frame ! Background Frame !
+!
-!
+!
-!
Abs(.) > ?

Abs(.) > ?

Update: Background[t] := Background[t-1] + (1- ) New[t]

Image subtraction example from IC manufacturing: die-to-die comparison of photomasks !
Where is the defect? !
Image A (no defect)!
Image B (w/ defect)!
Absolute difference between two images!
w/o alignment!
w/ alignment and thresholding!

Chapter Overview: Point Operations!: Relating Gray Values To Brightness!

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter Overview: Point Operations!: Relating Gray Values To Brightness!

Uploaded by

Copyright:

Available Formats

Chapter Overview: Point Operations !

Relating gray values to brightness!

Averaging! Subtraction! The need for image registration!

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 1 !

Quantization: how many bits per pixel? !

How many gray levels are required? !

Contouring is most visible for a ramp!

32 levels! 64 levels! 128 levels! 256 levels!

Digital images typically are quantized to 256 gray levels. !

Brightness discrimination experiment !

Can you see the circle?!

125! 126! 128! 129! 130! 127!

Note: I is luminance, measured in cd m 2

Weber fraction! Webers Law!

Contrast ratio without contouring !

Luminance ratio between two successive quantization levels at visibility threshold!

I max N 1 = (1 + KWeber ) I min

For! KWeber = 0.010.02 N = 256 Typical display contrast ratio!

I max = 13156 I min

Suggests uniform perception in the log(I) domain (Fechners Law)!

Cathode ray tubes (CRTs) are nonlinear!

Cameras contain -predistortion circuit!

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 6 !

log vs. -predistortion !

Similar enough for most practical applications!

Lower speed lms tend to have higher absolute !

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 8 !

Brightness adjustment by intensity scaling !

Scaling in the -domain is equivalent to scaling in the linear luminance domain

. . . same effect as adjusting camera exposure time. !

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 9 !

Contrast adjustment by changing !

. . . same effect as using a different photographic lm . . .!

Contrast adjustment by changing !

Scaled ramp 20!

Scaling chosen to! approximately preserve! brightness of mid-gray!

Scaled ramp 0.50!

Bernd Girod: EE368 Digital Image Processing!

Gray level histograms !

Distribution of gray levels can be judged by measuring a histogram:!

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 12 !

Cameraman ! image! gray level!

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 13 !

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 14 !

Idea: nd a non-linear transformation!

Goal: pdf pg(g) = 1 over the range ! g 1 0

Bernd Girod: EE368 Digital Image Processing!

Histogram equalization for continuous case !

From basic probability theory!

Consider the transformation function!

Bernd Girod: EE368 Digital Image Processing!

Point Operations no. 16 !

Histogram equalization for discrete case !

Discrete approximation of! g = T ( f ) = p f ( ) d

Bernd Girod: EE368 Digital Image Processing!