PowerPoint Presentation by X0kg4Mt

VIEWS: 0 PAGES: 10

• pg 1
```									                 BTC Encoding

S.R.Subramanya                  1
Block Truncation Coding(in:I; out;code)

• Divide the image into non-overlapping blocks of n x ni pixels
• For each block
› Compute the mean,  ,and standard deviation, ,for each
block   12 n1 P[i , j ] ;   12 n n P 2 [i , j ]   2 where P[i,j]
i
n                            j 1   j 1
n
is the value of pixel at (i,j).
› Detemine p and q be the number of pixels in the block
which have values greater or equal, and less than 

S.R.Subramanya                                                             2
Block Truncation Coding(in:I; out;code)

›Compute reconstruction levels a and b
a    p q , b    q p

› Quantize the pixels in each block to get a bitmap whose
pixel values are; 1 if P[i, j ]   and 0 otherwise

• Encode the bitmaps and reconstruction levels

S.R.Subramanya                                              3
BTC Decoding

Decode
bitmap
Form      Combine
recons.   blocks
Decode     block
recons.
values
Reconstructed
Image

S.R.Subramanya                                       4
CCC Encoding

reconstructed
values
a’

each        b’
block

Original Image            bitmap

color lookup table
S.R.Subramanya                                             5
Color Cell Compression

NM            Dimensions of the original image I
n m           Dimensions of each block
P (i , j )     Pixel value in a block at relative position (i,j)
PR , PG , PB The (R,G,B) components of pixel P(i,j)
Y (i , j )      Luminance (brightness) components of pixel P(i,j)
              Mean brightness of a block
aC , bC        Reconstruction pixel values
p(q )
S.R.Subramanya
Number of pixels greater than,equal to (less than)
6
Color Cell Compression(in: I; out; code)

• Divide the image into non-overlapping blocks of n m
pixels. For each block do the following
• For each pixel P(i, j ) compute the luminance (brightness)
component, Y(i,j):
Y (i , j )  0.3 PR (i , j )  0.59 PG (i , j )  0.11PB (i , j )

• Compute the mean brightness,  :
1
          i 1  j 1 Y ( i , j )
n     m

nm

• Compute reconstruction values aC and bC ,where C {R, G, B}
1                                                1
aC        
p i , j:Y ( i , j ) 
PC ( i , j )   ,   bC      
q i , j ;Y ( i , j ) 
PC ( i , j )
S.R.Subramanya                                                                                            7
Color Cell Compression(in: I; out; code)

• Compute bitmap B:
B( i , j )  1 ,if Y ( i , j )   ;   B( i , j )  0   ,if   Y (i , j )   ;

• Quantize (aR , aG , aB ) and (bR , bG , bB ) using CLT

• Let the corresponding indices into the CLT be a ' and b '

• Encode B, a ', b ' using entropy coding

S.R.Subramanya                                                                          8
Lossy DPCM Encoder

S.R.Subramanya                        9
Lossy DPCM Decoder

+
Sm
rm
Pm

S.R.Subramanya                        10

```
To top