International Journal of Computer Science Engineering and Information Technology Research (IJCSEITR) Vol.1, Issue 2 Dec 2011 169-178 TJPRC Pvt. Ltd., VIDEO COMPRESSION SYSTEM BASED ON JOINT PREDICTIVE CODING Asst. Prof. T. ARUMUGA MARIA DEVI 1, Ms. C. EBEN EXCELINE 2, Ms. K.K SHERIN 3 and Mr. MARIADAS RONNIE C.P 4 Assistant Professor, Dept. of CITE Assistant Professor PG Scholar PG Scholar Centre for Information Technology and Engineering, Manonmaniam Sundaranar University, Tirunelveli 1 Email: deviececit@gmail.com.phone No:9677996337 2 Email: ebencse@gmail.com Phone No: 9751387860 3 Email: sherinkk83@yahoo.com.phone No:9442055500 4 Email: mariadasronnie@yahoo.co.in.phone No:8089713017 ABSTRACT The Ultimate goal of video coding techniques is to store or transmit a video with less space or bandwidth. Video coding attempts to code each frame of the video which is temporally and spatially sampled. Thus video coding techniques are oriented towards construction of frames from the video, then code each frame and combine the coded frame to form a single video file. Video coding scheme could be used to video conferencing, mobile multimedia applications and for standard definition television and DVD video. Depending On the application, the compression algorithm is preferred. In particular, a novel approach that performs joint spatial and temporal prediction simultaneously is introduced. It bypasses the complex H.26x inter frame techniques and it is less computationally intensive. Because of the advantage of the effective joint prediction, the proposed approach is demonstrated experimentally to consistently lead to improved video quality, and in many cases to better compression rates and improved computational speed.
Asst.Prof. T. Arumuga Maria Devi, Ms. C. Eben Exceline, 170 Ms. K.K Sherin and Mr. Mariadas Ronnie C.P KEYWORDS: H.264/AVC, APT, Interpolative Coding, JPC I. INTRODUCTION The proposed video coding scheme reduces the computation complexity of the H.264 inter prediction and mode selection calculations by substituting it with a method based on interpolative prediction framework called Joint Predictive Coding [1] (JPC). It follows a new coding procedure that starts with an image color space transformation, followed by a pixel-to-pixel interpolative prediction, and ends in entropy coding in the spatial domain. The proposed framework introduces a new way of prediction that takes both temporal and spatial correlation into consideration simultaneously. Since, as mentioned above, efforts on pursuing accurate ME [7] seem to have reached their limit in increasing the compression rate, the proposed prediction method does not follow conventional methods. It performs intra prediction using Adaptive Prediction (Robinson [2]) but not in each frame separately, but instead in both the current and reference frame. The pixels in both frames are linked by MPEG like motion estimation. The calculated error of the reference frame is then used to estimate the intra prediction error in the current frame, and, using these values determines the joint predicted value. JPC has at least three advantages in comparison to the traditional scheme. 1. It utilizes both temporal and spatial correlation for prediction of each pixel, which leads to more accurate prediction, smaller error, and makes better compression. 2. The encoding is done on a pixel-to-pixel basis instead of on a block-toblock basis in traditional coding, avoiding visually sensitive blocking effects. Moreover, it tends to have good quality on videos with complicated textures and irregular contours, which is a difficult problem for block based codecs.
171 Video Compression System based on Joint Predictive Coding 3. Interpolative coding allows video frames to be reconstructed with increasing pixel accuracy or spatial resolution. This feature allows the reconstruction of video frames with different resolutions and pixel accuracy, as needed or desired, for different target devices. II. EXISTING METHODS FOR VIDEO COMPRESSION Most video coding schemes build upon the motion estimation (ME) and discrete cosine transform (DCT) frameworks popularized by ITU-T s H.26x family of coding standards. Although the current H.264/AVC [7] video coding standard achieves significantly better compression over its predecessors, namely, H.263, it still relies on traditional ME/DCT methods. By sacrificing computational time for high complexity motion estimation and mode selection, H.264 reduces the output bit-rate by 50% while increasing the encoding time significantly. Granted, the complex DCT of H.263 was abandoned for the simpler integer transform in H.264, however, the overall time spent calculating motion estimation is increased from 34% to 70% on average. This shows it is costly for H.264 to achieve current compression performance by pursuing accurate motion estimation. The compression efficiency is not likely to be improved significantly by modifying the ME module, and, consequently, the development potential of the ME/DCT [7] coding mechanism may be limited. III. FAST VIDEO COMPRESSION BASED ON JOINT PREDICTIVE CODING Preparation First, the proposed system computes the quantization intervals based upon a user-selectable quality parameter, which is derived in APT system. Then, it converts the video sequence in 4:4:4 format to number of video frames by temporal sampling.
Asst.Prof. T. Arumuga Maria Devi, Ms. C. Eben Exceline, 172 Ms. K.K Sherin and Mr. Mariadas Ronnie C.P Entropy coding Fn Color space transform Intra adaptive prediction Motion estimation Inter adaptive prediction F n-1 (reference) Color space transfor F n Inverse Color space transform Fig. 1 : Block Diagram for Video Compression based on Joint Predictive Coding Color space Transformation An RGB image may be converted to YCbCr after capture in order to reduce storage and/or transmission requirements. The equations for converting an RGB image to YCbCr color space are given in equations. Note that there is no need to specify a separate factor kg (because kb + kr + kg = 1) and that G can be extracted from the YCbCr representation by subtracting Cr and Cb from Y, demonstrating that it is not necessary to store or transmit a Cg component. Y = kr R + (1 kb kr )G + kbb Cb = 0.5 (B Y ) 1 kb Cr = 0.5 (R Y )
173 Video Compression System based on Joint Predictive Coding 1 kr Joint Predictive Coding Classical video coders choose between intra- and interprediction to remove the correlation in the video frames and, thus, achieve greater compression. Unfortunately, there are such cases that neither the temporal correlation nor the spatial correlation holds the dominant position. We propose a joint spatiotemporal prediction that consider both of them and lead to better accuracy in most circumstances. Intra prediction In intra prediction mode adaptive prediction trees method is used. In this method, pixels in an image are ordered in a series of groups called bands in increasing sample density and the encoder goes through all the bands from coarse to fine. The main idea is to use grids of points, oriented both squarely and diagonally, to predict their midpoints recursively. These groupings of square and diagonal points are alternately used for neighboring bands. Fig. 2 shows the band structure of interpolative coding. It represents a 5X5 image which is divided into bands denoted by letters a to e. Pixels in each band are predicted by those in higher bands. Pixels labeled a form a square with its center b1 ; thus,, and all equivalent bs throughout the image are predicted by as; similarly, Cs are predicted by as and bs, and so on. Different bands have different quantization intervals for the prediction errors. Pixels in primary bands, of more importance, have smaller quantization intervals, while pixels in latter bands with lower entropy are quantized coarsely. a1 e1 c1 e2 a2 e3 d1 e4 d2 e5 c2 e6 b1 e7 c3 e8 d3 e9 d4 e10 a3 e11 c4 e12 a4
Asst.Prof. T. Arumuga Maria Devi, Ms. C. Eben Exceline, 174 Ms. K.K Sherin and Mr. Mariadas Ronnie C.P Fig. 2. Band Structure U A B V X D C Fig. 3 : Predictive Grid In Fig. 3, each square represents a predictive grid. It helps to explain the algorithm of prediction. Pixel X is the center of the square grid (A,B,C,D). Pixels U and V are earlier processed neighbors in the same band as while pixels A,B,C,D belong to a higher band. To predict, bilinear prediction(a+b+c+d)/4 is usually ineffective. A nonlinear approach that selects a subset of its predictor set Px=(A,B,C,D,U,V) as predictors is proposed by Robinson inhis image compression scheme called Adaptive Prediction Trees (APT). We split the APT [2] prediction process into two parts. First, get the structure feature in terms of the order of amplitudes of the predictors. Represent this with a function Get Structure(), with its return value structure indicating which of the six possible pixels are used for prediction. Second, predict with the predictors based on the determined structure. This course is represented by a function Adaptive Pred(), which returns the estimate of. The prediction method in APT can be expressed as follows:
175 Video Compression System based on Joint Predictive Coding structure=get _structure(px) Xintra=Adaptive _pred(px,structure) Intra_error=X-Xintra The Get Structure() and Adaptive Pred() functions together not only determine which of the six neighboring pixels are to be used in the prediction, but specify how the prediction is calculated based upon the structure determined. Inter Prediction The JPC encoder does a motion vector search for every 16 X 16 block based on the primary component of the current frame. After this stage, each pixel has a motion vector (mv_x, mv_y). The matching pixel in the reference frame,x =(x+ mv_x,y+ mv_y), similarly, has its predictor set P x=(a,b,c,d,u,v ). Joint prediction takes both Px and P x to estimate. structure=get _structure(p x) X intra=adaptive _pred(px,structure) Inter_error =X -X intra Mode Selection JPC has two modes for blocks in video sequences of different features, Motion Compensation, and Intra.. Then, it pre predicts the current frame in Motion Compensation and Intra mode by mean square error. The mean square error (MSE) of the two modes in each 16 X 16 block, and are computed and the mode with smallest MSE is selected. Mode information and motion vectors are written into the bit-stream at the end of this stage. Entropy coding The residue frame is coded using adaptive Huffman coding. Huffman tree is initialized with a single node, known as the Not-Yet-Transmitted (NYT) or
Asst.Prof. T. Arumuga Maria Devi, Ms. C. Eben Exceline, 176 Ms. K.K Sherin and Mr. Mariadas Ronnie C.P escape code. This code will be sent every time that a new pixel value, which is not in the tree, is encountered. This allows for the decompressor to distinguish between a code and a pixel value. Also, the procedure creates a new node for the pixel value and a new NYT from the old NYT node. Whenever a pixel value that is already in the tree is encountered, the code is sent and the weight is increased. In order to for this algorithm to work, we need to add some additional information to the Huffman tree. In addition to each node having a weight, it will now also be assigned a unique node number. Also, all the nodes that have the same weight are said to be in the same block. These node numbers will be assigned in such a way that: 1. A node with a higher weight will have a higher node number. 2. A parent node will always have a higher node number than its children. After a count is increased, the update procedure moves up the tree and inspects the ancestors of the node one at a time. It checks to make sure that the node have the highest node in its block, and if not, swaps it with the highest node number. It then increases the node weight and goes to the parent. It continues until it reaches the root node.
177 Video Compression System based on Joint Predictive Coding Fig. 4 : Original Video Frame Fig. 5 : Reconstructed Video IV. CONCLUSION JPC may be substituted for standard lossless and lossy video and image Compression. JPC can, therefore, function as a general-purpose video and image compressor, freeing the encoder user from making decisions about which coding format is appropriate for the video and image. JPC can substitute for MPEG, DAT, JPEG, GIF, JPEG 2000, JPEG-LS, and PNG without risk of substantial underperformance. The good performance of the proposed joint predictive coding shows that joint spatiotemporal prediction is a serious competitor to the current block based video coding techniques. With prediction error reduced by 24.6% on average compared to the traditional inter/intraprediction [4] approach, joint spatiotemporal prediction is obviously superior to the conventional inter/intraprediction. ACKNOWLEDGMENT
Asst.Prof. T. Arumuga Maria Devi, Ms. C. Eben Exceline, 178 Ms. K.K Sherin and Mr. Mariadas Ronnie C.P The authors would like to thank the members of the Dept. of CITE, M S University, Tirunelveli for various algorithms used in this research, and all the people who helped in preparing and carrying out the experiments and data collection. REFERENCES 1. Wenfei Jiang, Longin Jan Latecki A Video Coding Scheme Based on Joint Spatiotemporal and Adaptive Prediction IEEE Trans. Image processing, vol. 18, no. 5, may 2009 2. J. A. Robinson, Adaptive prediction trees for image compression, IEEE Trans. Image Process., vol. 15, no. 8, pp. 2131 2145, Aug. 2006. 3. W.Woods and S. D. O Neil, Subband coding of images, IEEE Trans. Acoust., Speech, Signal Process., vol. 34, no. 5, pp. 1278 1288, Oct. 1986. 4. M. G. Day and J. A. Robinson, Residue-free video coding with pixelwise adaptive spatio-temporal prediction, IET Image Process., vol. 2, no. 3, pp. 131 138, Jun. 2008 5. J. A. Robinson, A. Druet, and N. Gosset, Video compression with binary tree recursive motion estimation and binary tree residue coding, IEEE Trans. Image Process., vol. 9, no. 7, Jul. 2000. 6. E. A. Gifford, B. R. Hunt, and M. W. Marcellin, Image coding using adaptive recursive interpolative DPCM, IEEE Trans. Image Process., vol. 4, no. 8, pp. 1061 1069, Aug. 1995. 7. H.264 and MPEG-4 Video Compression Video Coding for Nextgeneration Multimedia Iain E. G. Richardson The Robert Gordon University, Aberdeen, UK.
179 Video Compression System based on Joint Predictive Coding 8. MPEG-2ReferenceSoftware, [Online].Available:http://www.mpeg.org/MSSG, MPEG-2 version 1.2, MPEG Software Simulation Group. 9. JM Software, JM 12.4, H.264/AVC Software Coordination, [Online]. Available:http://iphome.hhi.de/suehring/tml