Image Compression Using the 2D Wavelet Transform

1
Image Compression Using the 2-D Wavelet Transform
2
Introduction

• Other Image Compression methodDiscrete Cosine
  Transform(DCT),DPCM,the Entropy coding
• Removing RedundancyLimited amount of compression.

3
Wavelet Transform Steps

• Wavelet multiresolution image decomposition
• Quantization of the transform coefficients
• Encoding the quantized coefficients

4
Wavelet multiresolution image decomposition
GH
GG
HG
Hlow-pass filter Ghigh-pass filter
5
The Compression Algorithm

• The low-pass image HH is coded in full
• The remaining subbands are coded recursively by
  applying the following subroutine SendTree

6
Subroutine SendTree
Let Ir,s(x,y) be the coefficient value of the
decomposition I at octave r ,orientation s,
position x,y
7
Block Tree Structure
8
Factors in noise sensitivity of the eye

• Background luminance
• Spatial masking (Edge proximity)
• Band sensitivity
• Texture masking

9
Quantization of the transform coefficients
10
Encoding the quantized coefficientsHuffman Coding
First two bitsthe magnitude of the
coefficient The 3rd bitthe sign
11
Results

• Two standard monochrome test imageMiss America
  and Lena
• Compare with SNR values