Download as PPTX, PDF, TXT or read online from Scribd
Download as pptx, pdf, or txt
You are on page 1/ 19
JPEG (Joint Photographic Expert Group)
IMAGE COMPREESSION TECHNIQUE
Image Compression and it’s Types Objective of Image Compression is to reduce irrelevant and redundant image data in order to be ale to store or transmit data in efficient form
Lossless image compression is a compression algorithm that allows the original image to be perfectly reconstructed from the compressed data (PNG-where redundancy reduction plays a major role)
Lossy image compression is a type of compression where a certain amount of information is discarded, which means that some data are lost and hence the image cannot be perfectly reconstructed from the compressed data
JPEG comes under lossy compression, it deliberately losses information so that storage space reduction is more than 75 percent What sort of Info JPEG gets rid off?
Our eyes are more sensitive to brightness and also our eyes are less sensitive to high frequency components JPEG PROCESS ALGORITHM 01 02 03
Splitting RGB to YCbCr DCT
8x8 non overlapping pixels Colour treatment Frequency transformation (zero padding)
04 05 06
Quantization Lossless encoding Compressed Format
Based on quantization tables For further reduction File format with reduced size and JPEG standards RGB to YCbCr Chroma Subsampling (4:2) or Chroma Downsampling (4:2)
Y-64/64 - 100%
Cb-16/64- 25% Cr-16/64- 25%
50% reduction is achieved
FORWARD DCT The DCT uses the cosine function, therefore not interacting with complex numbers at all. DCT converts the information contained in a block(8x8) of pixels from spatial domain to the frequency domain.
Why DCT?
Human vision is insensitive to high frequency components, due to which it is possible to treat the data corresponding to high frequencies as redundant. To segregate the raw image data on the basis of frequency, it needs to be converted into frequency domain, which is the primary function of DCT. But the image matrix is a 2-D matrix. So we can either apply 1-D DCT to the image matrix twice. Once row-wise, then column wise, to get the DCT coefficients. Or we can apply the standard 2-D DCT formula for JPEG compression. If the input matrix is P(x,y) and the transformed matrix is F(u,v) or G(u,v) then the DCT for the 8 x 8 block is computed using the expression Considering Y pixels ENERGY COMPACTION (a property of DCT)
Most of the largest values
are concentrated in a few low frequency coefficients QUANTIZATION Removing high frequency components after the frequency separation done by DCT
STANDARD FORMULA: round ()
F(u, v) represents a DCT coefficient
Q(u, v) is a “quantization matrix”
What is Quantisation matrix?
Provided by JPEG standard from visual experiments. Basically a quality factor of 50 is
used ,increasing quality factor reduces the qunatisation table values. There is a separate table for Luma and Chrominance channel LUMINANCE CHANNEL CHROMINANCE CHANNEL DCT table Quantisation table Quantized table
/ =
Encoding with these matrix tend to eliminate the lower corner part to zero (eliminating higher frequency),while decoding with the same matrix only upper part can be recovered, hence it is a lossy compression RUN LENGTH ENCODING(ZIG ZAG SCAN)
Basically to reduce storage space occupied by redundant bits
DPCM ON DC COEFFICIENTS The DC coefficients are coded separately from the AC ones. Differential Pulse Code Modulation (DPCM) is the coding method.
If the DC coefficients for the first 5 image blocks are 150, 155, 149, 152, 144, then the DPCM would produce 150, 5, -6, 3, -8, assuming di = DCi+1 − DCi, and d0 = DC0.
HUFFMAN CODING ON DC AND AC COMPONENTS
The DC and AC coefficients finally undergo an entropy coding step to gain a possible further compression. Each DPCM coded DC coefficient or RLE AC coefficient is represented by (SIZE, AMPLITUDE), where SIZE indicates how many bits are needed for representing the coefficient, and AMPLITUDE contains the actual bits. 150, 5, −6, 3, −8 will be turned into (8, 10010110), (3, 101), (3, 001), (2, 11), (4, 0111) SIZE is Huffman coded since smaller SIZES occur much more often. MERITS • High Compression Ratio-Nearly 75% reduction in storgae • Color Support-Most of the compression doesn’t support color files • Variable Comression Levels-Based on quantisation tables,desired compression level can be achieved • Widespread Compatibilty-platform independent file
DEMERITS • Lossy Compression Artifacts- blockiness, blurriness, and color bleeding, especially at higher compression levels • Not Suitable for Text and Line Art: JPEG compression is optimized for photographic images and is not well-suited for images with sharp edges, text, or line artVariable
