DSpace at VNU: Image and video compression for wireless networks

DSpace at VNU: Image and video compression for wireless networks tài liệu, giáo án, bài giảng , luận văn, luận án, đồ án...

VNU JOURNAL OF SCIENCE, Mathematics - Physics, T.XXII, N01, 2006 IMAGE AND VIDEO COMPRESSION FOR WIRELESS NETWORKS Ho Anh Tuy Hanoi University o f Technology N guyen Vinh An Hanoi Open University Abstract The demands for transmission of image, video and multimedia over wireless networks are increasing very rapidly The inherently characteristics of wireless medium as bandwidth narrowing, interferences and fading are big challenges The traditional compression techniques can not adapt to this changes In this paper, we will discuss of issues of using wavelet compression for image and video over wireless networks This paper also introduces some change of wavelet parameters in order to improve quality of transmission image and video over wireless network conditions I n tr o d u c tio n The tra d i ti o n a l mobile networks are used for low - rate audio communications New generation of mobile comm unications is em e rgin g to t r a n s f e r d a ta trafic a t m uch higher bit rate The dem an ds for m ultimedia, im age a n d video over wireless netw ork s are rapidly increasing F u t u r e I n t e r n e t will allow u se rs access from mobile appliances (PDAs, s m a r t phone, Web pads, h a n d PCs ) and high b a n d w id th applications (m-commerces, m ultim edia E-mail, video telephone, wireless LANs, PANs) T r a n s m i s s io n of Image a nd Video over Wireless Networks a re challe nging because of th e highly variable n a t u r e of the wireless link Radio is a n i n h e r e n tly unreliable t r a n s m i s s io n medium when compared to a wired link due to inteference and fading c re a t e h igher bit errors Typical wireless c hannels are noisy a n d of n arrow b a n d w id th For example, a customer using a code-division multiple-access (CDMA) has only a 9.6 kbps bandw idth Even if the b a n d w id th i n cre as es up to Mbps for t h e 3G wireless, it is still not comparable to th e b a n d w id th of bro ad ba n d optical comm unication syste ms (ATM could allocate dozens of Mbps to end users) Time-varying c h ar acteristics of wireless channel a n d limited b a tt e r y resource in h an dh eld devices is a n o th e r issue for scalable video st r e a m in g over wireless link with Quality of Service (QoS) Meanwhile, the capacity of a wireless c h a n n e l is fluctuated d ue to th e changing distance between the t r a n s m i t t e r a nd t h e receiver So it is i m p o r t a n t to e stim a te the available wireless network condition dyna mically and it is n e ce ssa ry to apply appro priate d video compression stra te g y to h a n d le the variability of wireless networks 33 Ho Anh Tuy, Nguyen Vinh An 34 Im age c o m p r e ss io n t e c h n iq u e s There are four common methods for compression, Discreate Cosine T r a n sf o r m (DCT), Vector qu a n ti za tio n (VQ), Fra ctal Compression and Discrete Wavelet Transform (DWT) DCT is a lossy compression algorithm t h a t samples an image a t r e g u la r intervals, analyzes the frequency components p r e s e n t in the sample, discards those frequencies which not affect the image as the h u m a n eye perceives it Vector qu an tizatio n (VQ) is a lossy compression t h a t looks a t a n a r r a y of data, not an individual value It compresses r e d u n d a n t d a ta while a t the sam e time r etaining th e desired object or da ta s t r e a m ’s original intent Fractal compression is a form of VQ and is also a lossy compression The self­ similar sections of image is located and then using fractal algorithm to handle them DWT analyzes signals into wavelets-functions t h a t have both time and frequency domains The process is perform on the entire image, which differs from DCT t h a t works on sm aller block (8 x picxels) 2.1 D iscrete Cosine T r a n s fo r m (DCT) Currently, DCT is quite popular in m any compression products such as J P E G Image compression using DCT is illustrate d in figure F ig u r e JP E G compression The general forward a nd inverse DCT transfo rm for a 2D (N by M image) is defined by the following equation DCT: s EY \ _ r v w \ v l ' v l /Y \ ( 2x + X)vn {2y+\)nn F(w,v) = f - C ( t < ) C ( v ) X f ( y , x ) c o s -—^ - COS - — N y=0 x=0 N + \)U T (1) 2N IDCT: t M which can recover e rrors of length up to (N-M)/2.By a d d in g extra p a rity d a ta of length R bytes, a t least p a r t of errors can be recovered by th e receiver The larg e r the value of R, more th e errors will be corrected Choosing of R should be considered carefully because parity d a ta in tr o d uces more trafic to th e limited netw ork band w id th a nd may cause packet loss due to congestion The Peak Signal-to-Noise Ratio (PSNR) m e a s u r e s th e size of error relative to peak value of the signal Xpeak In other word, it is used to m e a s u r e the fidelity of a compressed image with its original High PSNR m e a n s t h a t the compressed image is very sim ila r to the original The for m ular to calculate PS NR is (6 ) where X peak is the peak value of the signal and Ơ2 * d is the Mean Square Error MSE I m a g e a n d Video C o m p re ssio n f o r Wireless N e t w o r k s 41 Mean s q u a r e error (MSE) m easu re s the difference betw een the original a nd reconstructed im age is calculated by 140% I o 120% w QJ N TO c 138%“ 135% % I 100% 100% ị a 100% E tfl c o _ Eo o z 80% 60% 40% ■ 40 % 20% 0% am level le ve l 25% C\J /0 level le ve l Wavelet Transform Level □ C o m p uta tio n E ne rgy ■ C o m m u n ic a tio n E nergy F ig u r e Effects of vary ing wavelet transfo rm level on energy consumption (computation and communication energy) 5.5.2 Varying quantizatio n level Effects of varying quantization level on image quality and communication energy — PSNR — 10 20 N o rm a lize d C o m m u n ica tio n E nergy 30 40 50 60 70 80 90 100 Quantization level F ig u r e Effects of communication energy va ry ing qu an tizatio n level on image quality and Im a g e a n d Video C o m p re ssio n f o r Wireless N e t w o r k s 45 The purpose of q u an tizatio n is to reduce th e entropy of the transformed coefficients so t h a t the t arg e t bit rate can be met Each of th e transfo rm coefficients a,, (u,v) of the subband b is quantized to the value q h (u ,v) according to the formula qh(u,v) = sign(ah(u,v)) M « v ) I (8) The qu an tizatio n step size A|, IS r e p r sented relative to the dynamic range of subband Each su b ba n d h as se parate qu an tizatio n step-size and only one quantizatio n step-size is allowed per subband [8 ], Varying the qua ntiza tio n level of the wavelet compression has several effects on mobile communication By increasing the qua n tiza tio n level, the n u m b er of t r a n s m i t te d bits will decrease, leading to a lower bit rate, less communication energy and ban d w id th required Of course, there is negativ e effect such as the image quality will degradation C o n clu sio n F u t u r e deployment of mobile multimedia d a ta servies a nd wireless video will require very large a m o u n t s of da ta to be tr an s m itte d However, transmiss io n of image a nd video over mobile and wireless network have some bottlenecks including limited ban dwidth , channel noise, b a ttery co nst rains of th e appliances This paper p res en t some Image a nd Video compresion popular techniques, some challenges of wireless netw orks for high bit rate data We propose some ways to improve the quality of tran s m iss io n of video and multimedia over wireless networks Based on wavelet image compression, we give some ideas to change p a r a m e t e r s of wavelet compression technique a t the source, which in tu r n will have to a d a p t to the verying wireless ne tw ork condition R e fe r e n c e s Clark N.Taylor and Sujit Dey, “Adaptive Image Compression for Wireless Multimedia Communication.”, Communications, ICC 2001 IEEE International Conference on , Volume: 6(2001), Page(s): 1925-1929 Qian Zhang, Wenwu Zhu a nd Ya-Quin Zhang, “Network-adaptive Scalable Video S t r e a m in g Over G Wireless Network”, Microsoft R esearch, China F Bejing Sigma Center, No 49, Zhichun Road Haidian District, Beijing 100080 p R China G a n g Ding, Hilima Ghafoor and B h a r a t Bhargava, “ E rro r Resilient Video T n sm iss io n over Wireless Ne tworks”, D e p a r t m e n t of Co m p u te r Sciences D e p a r t m e n t of Electrical and Computer Engineering, P u rdu e University West Lafayette, IN 47907 C.E Sh a nn o n, vol.37(1949) C om m unications in the Presence o f Noise Proc IRE 46 Ho Anh Tuy , Nguyen Vinh An R ak sh ith K r ish n a p p a,’’Image Compression Techniques and Video S t r e a m in g for Wireless M ultimedia Commun ication”, Illinois In stitu te o f Technology, ECES12, Ju ly 2003 Hua Zhu, Hao Wang, Imrich Chlamtac, Biao Chen, “B andw idth Scalable SourceChannel Coding for Video over Wireless Networks, www.utdallas.edu/~haowz/publication/ Dong-Gi Lee and Sujit Dey, “Adaptive a nd Energy Efficient Wavelet Image Compression For Mobile Multim edia D a ta Services”, Proc IE E E International Conference on C o m m u n ica tio n , April 2002 (called EEWITA) Athanassios Skodras, Ch arilaos Christopoulos a nd Touradj Ebra himi, “The J P E G 2000 Still Image Compression S t a n d a r d ”, I E E E Signal Processing Magazine, 2001  ... bit rates to very high MPEG-4 is a new generation of In tern et- b ased video applications a nd Video Coding Experts Group H.264 s t a n d a r d s for video compression is now widely used in videoconferencing... verying wireless ne tw ork condition R e fe r e n c e s Clark N.Taylor and Sujit Dey, “Adaptive Image Compression for Wireless Multimedia Communication.”, Communications, ICC 2001 IEEE International... rate M P E G -1 was designed for up to 1.5 Mbps, standad ized for compression of moving pictures a n d audio MPEG-2 was designed for between 1.5 and 15 Mbps It is based on M P E G - , b u t for