IMPLEMENTATION AND PERFOMANCE ANALYSIS OF H.264 INTRA FRAME CODING, JPEG, JPEG-LS, JPEG-2000 AND JPEG-XR 1 EE 5359 Multimedia Project Amee Solanki (1000740226)

IMPLEMENTATION AND PERFOMANCE ANALYSIS OF H.264 INTRA FRAME CODING, JPEG, JPEG-LS, JPEG-2000 AND JPEG-XR

1

EE 5359 Multimedia ProjectAmee Solanki (1000740226)amee.solanki@mavs.uta.edu

Image Compression

2

•Compression is the process of compacting data, reducing the number of bits. •Reduce redundancy of the image or video data in order to be able to store or transmit data in an efficient form.

Fig.1 Comparison of original coronary angiogram (left) with two compression results. Middle: JPEG data compression by factor of CR=12, Right: factor of CR=24[14].

Two Types of Compression

3

Lossless compression: There is no information loss, and the image can be

reconstructed exactly the same as the original Applications: medical imagery, archiving

Lossy compression: Information loss is tolerable. Applications: commercial distribution (DVD) and rate

constrained environment where lossless methods cannot provide enough compression ratio

Evolution of Image Compression Standards

4

Fig.2 Evolution of compression technology[15]

Compression standards

5

Standard Software Main Application

Year

JPEG JPEG-Baseline Ref.

Image 1992-1999

JPEG-LS JPEG-LS DLL*DLL-Dynamic linked library

Image 1999-2000

JPEG-2000 JasPer Image 2000

JPEG-XR JPEG-XR Ref. Image 2009

H.264/AVC Intra Coding

JM Video 2003

Table 1: Comparison of image compression standards[13]

JPEG Standards

6

Baseline JPEG Encoder and Decoder

7

Fig.2 JPEG encoder block diagram [1]

Fig.3 JPEG decoder block diagram [1]

JPEG 2000 Encoder and Decoder

8

Fig. 4 (a) Encoder block diagram (b) Decoder block diagram of JPEG 2000 [2]

JPEG and JPEG-2000

9

Standard

Compression ratio

Main CompressionTechnologies

Main TargetApplications

JPEG Compression ratio 2-30

-DCT-Perceptual quantization-Zig zag reordering-Huffman coding-Arithmetic coding

-Internet imaging-Digital photography-Image and video editing

JPEG-2000

Compression ratio 2-50

-Wavelets EBCOT

-Internet imaging-Digital photography-Image and video Editing-Printing-Medical imaging-Mobile applications-Color fax-Satellite imaging

Table 2: Comparison of JPEG and JPEG 2000 [13]

JPEG-LS and JPEG-XR

10

Standard Compression ratio

Main CompressionTechnologies

Main TargetApplications

JPEG-LS Compression Ratio 2:1

-Context Modeling-Prediction-Golomb Codes -Arithmetic coding

- Lossless and near lossless coding of continuous tone still images

JPEG-XR Higher compression ratio than JPEG

Based on HD Photo of Microsoft (Windows Media Photo)

-Storage and interchange of continuous tone photographic content (lossless and lossy )

Table 3: Comparison of JPEG-LS and JPEG-XR [13]

H.264/AVC(Advanced Video Coding) Standard

11

H.264 Basics

12

H.264/AVC compression video coding is based on the traditional hybrid concept of block-based motion-compensated prediction (MCP) and transform coding

In order to improve the compression efficiency of intra-only compression, the following two coding tools provide major contributions to the significant bit rate savings:

• Entropy encoding improvement, CAVLC and CABAC• Spatial intra prediction conducted by using spatially

neighboring samples of a target block which have been previously coded.

13

Fig.4 Examples of spatial intra prediction modes for (8X8) blocks [15]

Spatial Intra prediction[15]

•H.264/AVC uses both spatial and temporal predictions to increase its coding gain. •The intra-only compression uses spatial prediction and the prediction only occurs within a slice

14

Fig. 5 A 4X4 block and its neighboring pixels[16]

Fig. 6 Direction of 9 4X4 intra-prediction [16]

•Fig. 5 shows a 4x4 block containing 16 pixels labeled from a through p. A prediction block p is calculated based on the pixels labeled A-M obtained from the neighboring blocks.

•A prediction mode is a way to generate these 16 predictive pixel values using some or all of the neighboring pixels in nine different directions as shown in Fig. 6.

•In some cases, not all of the samples A-M are available within the current slice.

•In order to preserve independent decoding of slices, only samples within the current slice are used for prediction.

15

Fig.7 Examples of spatial intra prediction modes for (4X4) blocks[16]

1. Mode 0 is the vertical prediction mode in which pixels a, e, i, and m are predicted by A and so on.

2. Mode 1 is the horizontal prediction mode in which pixels a,b, c, and d are predicted by I and so on.

3. Mode 2 is called DC prediction in which all pixels i.e. (a to p) as shown in fig. 5 are predicted by (A+B+C+D+I+J+K+L)/8.

4. For modes 3-8, the predicted samples are formed from a weighted average of the prediction samples A-M.

H.264 Basic Encoder and Decoder

16Fig.8 (a) H.264 encoder block diagrams [3]

17

Fig.8 (b) H.264 decoder block diagrams [3]

Compressed Image Quality Measures

18

Criteria to evaluate a compressed image are as follows :

1. Compression ratio2. Bit-rate (bandwidth)3. Objective quality measure- PSNR, MSE (quality

of compressed image)4. Structural quality measure- SSIM

PSNR and MSE

19

Peak signal-to-noise ratio, often abbreviated PSNR, is the ratio between the maximum possible power of a signal and the power of corrupting noise that affects the fidelity of its representation

MSE and PSNR for a NxM pixel image are defined as

(1)

(2)

where x is the original image and y is the reconstructed image. M and N are the width and height of an image and ‘L’ is the maximum pixel value in the NxM pixel image

Structural Similarity Index

20

The structural similarity (SSIM) [17] index is a method for measuring the similarity between two images

SSIM is designed to improve on traditional methods like peak signal-to-noise ratio (PSNR) and mean squared error (MSE), which have proved to be inconsistent with human eye perception

SSIM considers image degradation as perceived change in structural information. Structural information is the idea that the pixels have strong inter-dependencies especially when they are spatially close

SSIM Metric [17]

21

wherex and y correspond to two different signals that need to be compared, i.e. two different blocks in two separate images ;

Example with SSIM index

22

Fig. 9 SSIM Index example [4]

Conclusion

23

24

25

TABLE OF ACRONYMS

26

AVC advanced video codingBMPbit map formatCABAC context adaptive binary arithmetic codingDCTdiscrete cosine transformEBCOT embedded block coding with optimized truncationFRExt fidelity range extensionsGIF graphics interchange formatHD-photo high-definition photoHVShuman visual systemI-frame intra frameJMjoint modelJPEG joint photographic experts groupJPEG-LS joint photographic experts group lossless and lossless

codingJPEG-XR joint photographic experts group extended rangeLBT lapped bi-orthogonal transformLOCO-I low complexity lossless compression for imagesMSEmean square errorPSNR peak signal to noise ratioSSIM structural similarity index VLC variable length coding

References

27

[1] JPEG encoder and decoder block diagram :http://www.cmlab.csie.ntu.edu.tw/cml/dsp/training/coding/jpeg/jpeg/decoder.htm

[2] JPEG2000 encoder and decoder block diagram :http://eeweb.poly.edu/~yao/EE3414/JPEG.pdf

[3] H.264 encoder and decoder block diagram :S. Kwon, A. Tamhankar and K.R. Rao, “Overview of H.264 / MPEG-4 Part 10”, J. Visual Communication and Image Representation, vol. 17, pp.186-216, April 2006.

[4] SSIM index example diagram:https://ece.uwaterloo.ca/~z70wang/research/ssim/

[5] H.264/AVC reference software (JM 17.2) website: http://iphome.hhi.de/suehring/tml/download/

[6] JPEG2000 latest reference software (Jasper Version 1.900.0) website: http://www.ece.uvic.ca/~mdadams/jasper/

[7] JPEG reference software website: ftp://ftp.simtel.net/pub/simtelnet/msdos/graphics/jpegsr6.zip

[8] JPEG-LS reference software website: http://www.hpl.hp.com/loco/

28

[9] T. Wiegand, G. J. Sullivan, G. Bjontegaard and A. Luthra,” Overview of the H.264 / AVC video coding standard ” IEEE Trans. on Circuits and Systems for Video Technology,vol. 13, pp. 560-576, July 2003.

[10] A.Skodras, C. Christopoulos and T. Ebrahimi, “The JPEG 2000 still image compression standard”, IEEE Trans. on Signal Processing, vol.18, pp. 36 - 58, Aug 2002.

[11] M. J. Weinberger, G. Seroussi and G. Sapiro, “The LOCO-I lossless image compression algorithm: principles and standardization into JPEG-LS”, IEEE Trans. on Image Processing, vol.9, pp.1309-1324, Aug. 2000

[12] C. Christopoulos, A. Skodras and T.Ebrahimi, “The JPEG2000 still image coding system: anoverview”, IEEE Trans. on Consumer Electronics, vol.46, pp.1103-

1127, Nov. 2000.

[13] T. Ebrahimi and M. Kunt, “ Visual data compression for multimedia applications”, Proc IEEE, vol.86, pp. 1109-1125, June 1998.

29

[14] Image compression test image:

http://www.uni-kiel.de/Kardiologie/dicom/1999/compression1.html.

[15] Evolution of image compression standards :ftp://ftp.panasonic.com/pub/panasonic/drivers/PBTS/papers/WP_AVC-Intra.pdf

[16] Intra-prediction modes image:http://www.atc-labs.com/technology/h264_publication_1.pdf [17] Z. Wang, A. C. Bovik, H. R. Sheikh and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,” IEEE Trans. on Image Processing, vol. 13, no. 4, pp. 600-612, Apr. 2004.

[18] I. E. Richardson, “The H.264 advanced video compression standard”, II Edition, Wiley, 2010.

[19] D. S. Taubman and M. W. Marcellin, "JPEG2000 – Image compression fundamentals, standards, and practice," Kluwer, 2001.

Some Important terms

30

Quality factor –N denotes the scale quantization tables to adjust image quality. Quality factor varies from 0 (worst) to 100 (best); default is 75.

Rate specify target rate as a positive real number. ‘rate’=1 corresponds to no compression. Rate and bits per pixel are related by the expression: compression ratio=24/bpp= 1/rate for a color image and rate=bpp/8 for a gray scale image.

Error value is varied from 1 to 60. Error value of zero corresponds to no compression.T1, T2, T3 are thresholds. While giving the settings the following condition need to be met. (Error value+1)<T1<T2<T3.

Results

31

32

33

34

35

36

37

38

39

40