ECE160 / CMPS182 MultimediaThe JPEG StandardObservations for JPEG Image CompressionMain Steps in JPEG Image CompressionBlock diagram for JPEG encoderSlide 6DCT on Image BlocksSlide 8QuantizationLuminance Quantization TableChrominance Quantization TableJPEG compression for a smooth image blockSlide 13JPEG compression for a textured image blockSlide 15Slide 16Run-length Coding (RLC) on AC coefficientsSlide 18DPCM on DC coefficientsSlide 20Entropy CodingFour Common JPEG ModesProgressive ModeSlide 24Hierarchical ModeEncoder and Decoder for Three-level Hierarchical JPEGSlide 27Slide 28Slide 29Encoder for Three-level Hierarchical JPEGDecoder for Three-level Hierarchical JPEGThe JPEG BitstreamThe JPEG2000 StandardThe JPEG2000 standardSlide 35Slide 36Properties of JPEG2000 Image CompressionMain Steps of JPEG2000 Image CompressionLayer Formation and RepresentationSlide 55Region of Interest CodingSlide 57Performance comparison for JPEG and JPEG2000Comparison of JPEG and JPEG2000.Slide 60The JPEG-LS StandardJBIG and JBIG-2: Bi-level Image Compression StandardsECE160Spring 2009 Lecture 9Image Compression Standards1ECE160 / CMPS182MultimediaLecture 9: Spring 2009Image Compression StandardsECE160Spring 2009 Lecture 9Image Compression Standards2The JPEG Standard•JPEG is an image compression standard that was developed by the “Joint Photographic Experts Group". JPEG was formally accepted as an international standard in 1992.•JPEG is a lossy image compression method. It employs a transform coding method using the DCT (Discrete Cosine Transform).•An image is a function of i and j (conventionally x and y) in the spatial domain.•The 2D DCT is used as one step in JPEG in order to yield a frequency response which is a function F(u,v) in the spatial frequency domain, indexed by two integers u and v.ECE160Spring 2009 Lecture 9Image Compression Standards3Observations for JPEG Image Compression•Observation 1: Useful image contents change relatively slowly across the image. It is unusual for intensity to vary widely several times in a small area, for example, within an 8x8 pixel block.–Much of the information in an image is repeated, hence “spatial redundancy".•Observation 2: Psychophysical experiments show that humans are less likely to notice the loss of high spatial frequency components than the loss of lower frequency components.–The spatial redundancy can be reduced by reducing the high spatial frequency contents.•Observation 3: Visual acuity (accuracy in distinguishing closely spaced lines) is greater for gray (“black and white") than for color.–chroma subsampling (4:2:0) is used in JPEG.ECE160Spring 2009 Lecture 9Image Compression Standards4Main Steps in JPEG Image Compression•Transform RGB to YIQ or YUV and subsample color.•DCT on image blocks.•Quantization.•Zig-zag ordering and run-length encoding.•Entropy coding.ECE160Spring 2009 Lecture 9Image Compression Standards5Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards6Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards7DCT on Image Blocks•Each image is divided into 8 8 blocks. The 2D DCT is applied to each block image f(i,j), with output being the DCT coefficients F(u,v) for each block.•Using blocks has the effect of isolating each block from its neighboring context. This is why JPEG images look choppy (“blocky") when a high compression ratio is specified by the user.ECE160Spring 2009 Lecture 9Image Compression Standards8Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards9Quantization•F(u,v) represents a DCT coefficient, Q(u,v) is a “quantization matrix" entry, and F(u,v) represents the quantized DCT coefficients which JPEG will use in the succeeding entropy coding.–The quantization step is the main source for loss in JPEG compression.–The entries of Q(u,v) tend to have larger values towards the lower right corner. This introduces more loss at the higher spatial frequencies - a practice supported by Observations 1 and 2.–The Luminance Quantization Table and The Chrominance Quantization Table provide the default Q(u,v) values obtained from psychophysical studies to maximize the compression ratio while minimizing perceptual losses in JPEG images.ECE160Spring 2009 Lecture 9Image Compression Standards10Luminance Quantization TableECE160Spring 2009 Lecture 9Image Compression Standards11Chrominance Quantization TableECE160Spring 2009 Lecture 9Image Compression Standards12JPEG compressionfor a smooth image blockAn 8x8 block fromthe Y image of `Lena'ECE160Spring 2009 Lecture 9Image Compression Standards13JPEG compressionfor a smooth image blockECE160Spring 2009 Lecture 9Image Compression Standards14JPEG compressionfor a textured image blockAnother 8x8 block fromthe Y image of `Lena'ECE160Spring 2009 Lecture 9Image Compression Standards15JPEG compressionfor a textured image blockECE160Spring 2009 Lecture 9Image Compression Standards16Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards17Run-length Coding (RLC) on AC coefficients•RLC aims to turn the F(u,v) values into sets {#-zeros-to-skip , next non-zero value}.•To make a long run of zeros more likely, a zig-zag scan is used on the 8x8 matrix F(u,v)ECE160Spring 2009 Lecture 9Image Compression Standards18Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards19DPCM 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.ECE160Spring 2009 Lecture 9Image Compression Standards20Block diagram for JPEG encoderECE160Spring 2009 Lecture 9Image Compression Standards21Entropy Coding•The DC and AC coefficients finally undergo an entropy coding step to gain a possible further compression.•Use DC as an example: each DPCM coded DC coefficient is represented by (SIZE, AMPLITUDE), –SIZE indicates how many bits are used to represent the coefficient, –AMPLITUDE contains the actual bits.•In the example, codes 150, 5, −6, 3, −8 will become(8, 10010110), (3, 101), (3, 001), (2, 11), (4, 0111)