Comparative Analysis of Compression Techniques Used for Facsimile Data(FAX) A Krishna Kumar Department of Electronic and Communication Engineering, Vasavi College of Engineering, Osmania University, Hyderabad, India. akrishnakumar01@gmail.com Abstract: The various data compression techniques are classified into two types, Lossy and Lossless techniques, In this thesis a study of various lossless algorithms used for facsimile transmission are discussed. This work provides detail study about the CCITT group 3 and group techniques and provides a Matlab model for the compression algorithm for monochrome images. The compressed images will be of a compression ratio about 1:15 which will help in fast communication and computation of these images. A set of standard test images with different characteristics in size and dimension are compressed by the Matlab model implementing the CCITT group3 and group algorithms. The CCITT recommended international standard for Group 3 and Group algorithms are Modified Huffman (MH), Modified READ (MR) and Modified Modified READ (MMR) are implemented and compression factor and space saving are computed and a comparison of them is shown.the results achieved shows better compression is achieved using MMR than MH and MR. Keywords : FAX (Fascimile), GIF, CCITT, ITU, MH (Modified haffman, MR (Modified read), MMR (Modified Modified read). Introduction: In order to store digital image data in a pel-by-pel representation a huge amount of storage capacity is necessary. For example a DIN A page scanned with a resolution of 300 dpi (dots per inch) needs about 1.1 M byte storage capacity in binary representation. It deals with various methods to compress binary image data. The focus is to discuss their suitability for archiving usual business documents made up of text, graphics and halftone segments. Compression denotes compact representation of data. Examples for the kind of data you typically want to compress are e.g. Text, Source-code, arbitrary files, Images, Video, Audio data, Speech. The following example illustrates the need for compression of digital images. a. To store a colour image of a moderate size, e.g. 512 512 pixels, one needs 0.75 MB of disk space. b. A 35mm digital slide with a resolution of 12µm requires 18 MB. c. One second of digital PAL (Phase Alternation Line) video requires 27 MB. To store these images, and make them available over network (e.g. the internet), compression techniques are needed. Image compression addresses the problem of reducing the amount of data required to represent a digital image. Compression can be divided into two categories, (i) Lossless compression (ii) Lossy compression Figure : Tree representation of compression methods The lossless coding techniques are Run length encoding, Huffman encoding,arithmetic encoding, Entropy coding, Area coding,lzw coding. The lossy compression techniques are Predictive coding, Transform coding (FT/DCT/Wavelets). 1
The following evaluation criteria of a compression scanned in the binary mode. A scan line is a techniques are Compression ratio Compression ratio (cr)= number of input bits / number of output bits Storage space Space saved Space saved = before compression sizeafter compression size Compression efficiency Compression efficiency = 1- (1/cr) Existing system: Fax technical progress: 1876: telephone(bell) 1902: optical scan 1917: teletype(at&t) 1968: CCITT Group 1 1976: CCITT Group 2 1980: CCITT Group 3 198: Group 2003: 56 kbps fax/modem 1.5mbpsadsl/cable Fax communication block diagram: Figure :Fax communication block diagram Fax compression techniques: A binary image containing black and white pixels is generated when a document is complete line of pixels, of height equal to one pixel,running across the page. It scans the first line of pixels,the scans 2 nd line and works it s way to the bottom of the page,ending with the last scan line. Each scan line is scanned from the left of the page to the right of page generating black andwhite pixels for that scan line. The binary image compression methods are, Pack bits encoding CCITT Group 3 1D (MH) CCITT Group 3 2D (MR) CCITT Group 2D (MMR) MH (MODIFIED HUFFMAN ) The fax pages contain many runs of white and black pixels which make RLE efficient for minimizing these run lengths. The efficiently compressed run lengths are then combined with Huffman coding. Thus an efficient and simple algorithm is achieved by combining RLE with Huffman coding and this is known as Modified Huffman. RLE consists of terminating and makeup codes. MH coding uses specified tables for terminating and makeup codes. Terminating codes represent shorter runs while the makeup codes represent the longer runs. The white and black pixel runs from 0 to 63 are represented by terminating codes while greater than 63 are represented with makeup codes which mean that greater than 63 bit runs are defined in multiples of 6 bits which are formed by the terminating codes. These tables are given. A scan line represented with long runs gives a make code which is less than or equal to the pixel run and then the difference is given by the terminating code. The following example will help in understanding how it works. 2
There are three different types of bit pattern in MH Compression ratio(cr) = original size / compressed coding Pixel information (data) Fill EOL size = 205286/1818 = 13.8316377 Space saved = original size compressed size = 205286 1818 =1906 bits Compression efficiency = 1- (1/cr) =1-0.07229802 = 0.92770198 for input image 2: Consider the following image as input Figure :MH structure Table : Makeup codes for input image 1: Case studies: Consider the following image as input Figure. Input 2 image of MH The input image2 will be compressed using MH technique. << compression bits for input 2,after compression :271270 bits Original size of the input image: 1188*1728 = 205286 bits Figure. Input 1 image of MH The input image1 will be compressed using MH technique. This process will give the following output (matlab results) << compression bits for input 1,after compression :1818 bits Original size of the input image: 1188*1728 = 205286 bits Compression ratio(cr) = original size / compressed size = 205286/271270 = 7.5676023 Space saved = original size compressed size = 205286 271270 =178159 bits Compression efficiency = 1- (1/cr) =1-0.1321222 = 0.86785778 for input image 3: Consider the following image as input 3
The input image will be compressed using MH technique. This process will give the following output. << compression bits for input,after compression :219751 bits Original size of the input image: 1188*1728 = 205286 bits Figure. Input 3 image of MH The input image3 will be compressed using MH technique. This process will give the following output. << compression bits for input 3,after compression :11727 bits Original size of the input image: 1188*1728 = 205286 bits Compression ratio(cr) = original size / compressed size = 205286/11727= 17.8202 Space saved = original size compressed size = 205286 11727 =193537 bits Compression efficiency = 1- (1/cr) =1-0.05720155 = 0.927985 for input image : Consider the following image as input. Figure. Input image of MH Compression ratio(cr) = original size / compressed size = 205286/219751= 9.3177319 Space saved = original size compressed size = 205286 219751 =1833113 bits Compression efficiency = 1- (1/cr) =1-0.1070606 = 0.8929539 MR (MODIFIED READ) The Group 3 (MR) facsimile (the "one minute facsimile") was recommended as 'Redundancy suppressive digital facsimile' in 1980. At this time, Japan took the lead, and twodimension encoding was examined by CCITT as the basic facsimile redundancy reduction method. Joint proposals were made by KDD (Kokusai Denshin Denwa Co., Ltd., present KDDI) and Nippon Telegraph and Telephone Public Corporation (now Nippon Telegraph and Telephone Corporation = NTT). The twodimension encoding MR (Modified READ (= Relative Element Address Designate)) method was adopted as the basis for an international standard (ITU-T recommendation T.) in 1980. Typical G3 facsimile equipment of that time is shown in figure. The MR method was extended to yield the MMR method; it was adopted as ITU-T recommendation T.6 in 198 and is the encoding method of the current G facsimile[3]. This recommendation was issued for 'For telematique
equipment and ISDN (Integrated Service Digital Network)'. Figure : passmode Figure :MR structure Figure :Group3 architecture Case studies: MH case study images are consider for MR also, as explained in detailed above(mh) with calculations. But the output values of MR techniques are different from MH. MMR (MODIFIED MODIFIED READ ) MMR coding scheme is an extension of the MR coding scheme which has the same mode of coding i.e. pass mode, horizontal mode and vertical mode. These coding modes are described on the basis of the changing element in the scan line and the reference line these changing elements are also described in The MR coding. Modes of Coding This coding scheme has three modes. These modes are defined on the basis of the difference between the changing elements of the scan line to the reference line. Pass mode Vertical mode Horizontal mode Figure. : vertical mode Figure : Horizontal mode Figure : coding steps for G2D MH case study images are consider for MMR also, as explained in detailed above(mh) with calculations. But the output values of MMR techniques are different from MH and MR. 5
RESULT The following graph shows the comparison of The fax pages contain many runs of white and black pixels which make RLE efficient for minimizing these run lengths. The efficiently compressed run lengths are then combined with Huffman coding. Thus an efficient and simple algorithm is achieved by combining RLE with Huffman coding and this is known as Modified Huffman. compressed bits for MH, MR, MMR for images. four MR is also known as Modified Relative Element address designated (READ). MR exploits the correlation between successive lines. It is known that two consecutive lines have a very high percentage of single pixel transition due to a very high resolution of the images. ITU-T Recommendation T.6 gives the Modified Modified Read or MMR encoding algorithm. MMR is an upgraded version of the MR. They are both 2-dimensional algorithms but MMR is an extended version of the 2-dimension. The fundamentals of MMR are same as MR except a few minor changes to the algorithm however the modes of MR i.e. pass mode, vertical mode and horizontal mode are the same for MMR encoding. The major change in the MMR with respect to MR is the K parameter. The following table shows the comparisons of FAX compression techniques. Compressed MH MR MMR bits for Image 1 1818 111051 98760 Image 2 271270 1375 89620 Image 3 11727 7311 5877 Image 219751 156818 135326 Table : Comparison of compressed bits for MH,MR,MMR Figure : Comparison of compressed bits for MH,MR,MMR The following table shows the space saving of the image sizes. Space saving= original size compressed size Space 0riginal MH MR MMR saved (bits) for size Imag e 1 205286 190 6 19181 3 19510 Imag e 2 205286 178159 191811 9 19632 Imag e 3 205286 19353 7 197975 0 19938 7 Imag e 205286 183311 3 18960 191753 8 Table : Comparison of space saving for MH,MR,MMR The following graph shows the comparison of space saving for MH, MR, MMR for four input images. 6
Figure : Comparison of compression ratios (CR) for MH,MR,MMR The following table shows the efficiency (%) of the input image sizes. Compression efficiency = 1-(1/CR) compression Figure : Comparison of space saving for MH,MR,MMR The following table shows the compression ratio (CR) of the input image sizes. compression ratio (CR)= Before compression bits/ After compression bits Compression MH MR MMR ratio(cr )for Image 1 13.83 18.8 20.78 Image 2 07.56 15.23 22.90 Image 3 17.8 28.07 35.10 Image 09.3 13.09 15.16 Table : Comparison of compression ratios (CR) for MH,MR,MMR The following graph shows the comparison of compression ratios(cr) for MH, MR, MMR for four input images. Compression MH MR MMR efficiency (%) Image 1 92 9 95 Image 2 86 93 95 Image 3 9 96 97 Image 89 92 93 Table : Comparison of compression efficiency(%) for MH,MR,MMR The following graph shows the comparison of compression efficiency (%) for MH, MR, MMR for four input images. Figure : Comparison of compression efficiency(%) for MH,MR,MMR Table :Comparisons of MH,MR,MMR 7
The compression techniques of the FAX better algorithms that focus on more improved technologies are having different compression ratios. Finally the MMR technique having the best compression efficiency than MH and MR. colored images. It can be easily extended to gray images used by bit-plane slicing and color images used by MRC techniques. CONCLUSION The results taken from the MATLAB model suggests that this algorithm has the ability to compress images of any size to much lesser than the original size. Nowadays these algorithms are very popular with sensor based smart camera. These smart cameras are usually working in a network with a wireless node or connected in some remote location with a smaller data rate line. Hence the compression ratios of these algorithms makes them a popular choice and have re evolved a technology that has been sidelined by so many newer and better algorithms that focus on more improved coloured images. This study provides a base for implementation of this algorithm on any flexible hardware platform such as FPGA, micro controller etc. which will provide faster compression of image received through any medium such as cameras and scanners. FUTURE SCOPE However the complexity can be traded when Group algorithms are implemented on hardware. Nowadays these algorithms are very popular with sensor based smart camera. These smart cameras are usually working in a network with a wireless node or connected in some remote location with a smaller data rate line. Hence the compression ratios of these algorithms makes them a popular choice and have re evolved a technology that has been sidelined by so many newer and REFERENCES [1]. R. C. Gonzalez & R. E. Woods, Digital Image Processing, Addison Wesley 1992. [2 ] MATLAB User's Guide, The Math Works Inc., 1993. [3]. McConnell, FAX: Digital Facsimile Technology & Applications, Artech House, Norwood, MA; second edition, 1992. []. Marking, P. Michael, "Decoding Group 3 Images," The C Users Journal, June1990. [5]. R. Hunter and A.H. Robinson, "International Digital Facsimile Coding Standards," Proceedings of the IEEE, vol. 68, no. 7, July 1980. [6].Anton Kruger, "Huffman Data Compression," C Gazette, vol. 5, no., 1991. [7]. Research and Events that Permitted Facsimile Use to Explode in Japan by Kaoru Wakabayashi published in IEEE,2009. [8]. Azam Khan, Algorithm study and MATLAB model for CCITT Group TIFF Image Compression, At Linköping Institute of Technology,December 2010. [9]. Gleb V. Tcheslavski, Basic Image compression methods, spring 2008. [10]. Paula Aguilera, Comparison of different image compression formats ECE 533 Project Report. [11]. Kaoru Wakabayashi, Research and Events that Permitted Facsimile Use to Explode in Japan, NTT Cyber Space Laboratories, NTT Corporation, japan. [12]. Ann Cavoukian, Ph.D., Guidelines on Facsimile Transmission Security January 2003. 8
9