Efficient Spatial Averaging Filter for High Quality Compressed Wireless Image Transmission

Transcription

1 Efficient Spatial Averaging Filter for High Quality Compressed Wireless Image Transmission Santha Devi.P, Sivakumar. M & Arthanariee A 1 Mother Teresa Women s University, Kodaikanal, Tamil Nadu, India. Anna University, Coimbatore,Tamil Nadu, India. 3 Department of Science and Humanities, Nehru Institute of Technology, Kaliyapalayam, Coimbatore 105, Tamil Nadu, India. psanthabaskar@gmail.com, sivala@gmail.com, arthanarimsvc@gmail.com Abstract - In Wireless Multimedia Sensor Network (WMSN), image communication plays an important role to handle the issues, resource consumption and the quality of the image being transmitted without noise in wireless channels. Initially, the image is compressed by improved polyomines lossless compression technique, which increases the quality of image at receiving end of the wireless communication. Compressed image is transmitted by Energy Efficient High Quality Image Transmission scheme (EEHQIT) to achieve energy efficient image transmissions in WSNs. EEHQIT scheme is compelling due to its ability of saving individual power consumption over multiple sensors by spreading total transmission consumption. In this paper, spatial average filtering technique is proposed, removes the noise from the transmitted image and rebuilding the image to obtain the original image without loosing any portion of information from it. The spatial averaging filter was implemented, and tested on the transmitted compressed image. The experimental results show that the spatial averaging filter improves the image quality of the compressed image at the receiver side. Simulation results show up to 85% reduction in the total power consumption achieved and higher PSNR value achieved with the proposed noise filtering strategy. Keywords: Wireless Communication, Image Compression, Image quality, Interactive Transmission I. INTRODUCTION Several energy efficient protocols of image compression are proposed for wireless applications. The growth of 3G wireless communication systems in line with internet popularity made wireless multimedia image communication an important research topic in current network communication field. The characteristic of wireless multimedia communication which can be used to overcome the bandwidth and energy bottlenecks is that the conditions and requirements for mobile communication vary. ariations in wireless channel conditions may be due to user mobility, changing terrain, etc. The usual method for transmitting images over the Internet is to first compress the images using a lossy scheme such as JPEG, and then to transmit them across the intrinsically lossy Internet using the lossless TCP/IP protocol. JPEG and related lossy schemes are very sensitive to bit errors and hence require lossless transmission. The price paid for lossless transmission over a lossy medium is excessively lengthy transmission times due to retransmissions of lost packets. A more efficient means of transmitting the data is via some form of redundant transmission (forward error correction) which will make serious transmission errors unlikely. Redundancy must be applied selectively, however, since the addition of redundancy increases the amount of information to be transmitted. The compressed image is transmitted using EEHQIT scheme. Each imaging system suffers with a common problem of Noise. Unwanted data which may reduce the contrast deteriorating the shape or size of objects in the image and blurring of edges or dilution of fine details in the image may be termed as noise. It may be introduced by the image formation process, image recording, image transmission, etc. These random distortions make it difficult to perform any required picture processing. For example, the feature oriented enhancement is very effective in restoring blurry images, but it can be "frozen" by an oscillatory noise component. Even a ISSN (PRINT) : , olume -1, Issue -4,

2 small amount of noise is harmful when high accuracy is required, e.g. as in sub cell (subpixel) image analysisin this paper, we propose to denoise images by filtering the image. The Box filter is used for removing the noise from the image. II. RELATED WORKS Literature survey in [1][][4] addressed various issues regarding the challenges faced by research community in realizing WMSN. The early research efforts in wireless sensor networks did not investigate the issues of node collaboration, focusing more on issues in the design and packaging of small, wireless devices [5], more recent efforts (e.g. [6], [7]) have considered node collaboration issues such as data aggregation or fusion. Our approach of distributed image compression falls within the domain of techniques that apply the concept of in-network processing, i.e. processing in the network by computing over the data as it flows through the nodes. It is worth noting that current aggregation functions (e.g., maximum and average [7]) are limited to scalar data. Our approach can be viewed as an extension to vector data aggregation. Previous distributed signal processing/compression problems (e.g. [8], [9]) exploit correlations between data at close-by sensors in order to jointly compress or fuse the correlated information resulting in savings in communication energy. In parallel distributed computing theory [10], a problem (or task) is divided into multiple sub-problems (or sub-tasks) of smaller size (in terms of resource requirements). Every node solves each sub problem by running the same local algorithm, and the solution to the original problem is obtained by combining the outputs from the different nodes. Our approach to the design of distributed image compression is similar in concept, in that we distribute the task of image encoding/compression to multiple smaller image encoding/compression sub-tasks. our proposed approach of image compression intersects with the literature on lossy and lossless compression, which primarily focuses on polyomino technique. Digital images are prone to a variety of types of noise. There are several ways that noise can be introduced into an image, depending on how the image is created. For example: If the image is scanned from a photograph made on film, the film grain is a source of noise [11]. Noise can also be the result of damage to the film, or be introduced by the scanner itself. If the image is acquired directly in a digital format, the mechanism for gathering the data (such as a CCD detector) can introduce noise. Electronic transmission of image data can introduce noise [1]. Image noise elimination (reduction) [13] is the process of removing noise from the image. Noise reduction techniques are conceptually very similar regardless of the signal being processed [14], however a priori knowledge of the characteristics of an expected signal can mean the implementations of these techniques vary greatly depending on the type of signal. In practice a lot of methods are used to eliminate the noise from the image and a lot of filters are used. In this paper we used box filter for removing the noise from the image III. IMAGE TRANSMISSION USING BOX FILTER IN WMSN 3.1 Lossless Image Compression using Improved Polyomino There are two types of image compression: lossless and lossy. After decompression the original image is recovered. Compressing an image is significantly different than compressing raw binary data. The general purpose compression is used to compress images, but the result is less than optimal. This is because images have certain statistical properties which can be exploited by encoders specifically designed for them. This also means that lossy compression techniques can be used in this area. An integer-to-integer wavelet transform produces an integer-valued transform from the greyscale, integer-valued image. Since n loops in Bit-plane encoding reduces the quantization error to less than T0/ n, it follows that once n is greater than T0, and there will be zero error. In other words, the bit-plane encoded transform will be exactly the same as the original wavelet transform, hence lossless encoding is achieved Lossless compression involves with compressing data which, when decompressed, will be an exact replica of the original data. This is the case when binary data such as executables, documents etc. are compressed. They need to be exactly reproduced when decompressed. 3. Wavelet Image Compression Wavelet based image compression introduces no blocky artifacts in the decompressed image. It is decompressed image is much smoother and pleasant to eyes. We can also achieve much higher compression ratios regardless of the amount of compression achieved. By adding more and more detail information we can improve the quality. This feature is attractive for what is known as progressive transmission of images. Another lossy compression scheme developed for image compression is the fractal base image compression scheme (fig 1). However the fractal based image compression beginning to loss ground because it is very complex and time consuming. The filter components are reduced their size by half either by rejecting the even or odd samples thereby the 11 ISSN (PRINT) : , olume -1, Issue -4, 013

3 total size of the original signal is preserved. The low pass filter component retains almost all distinguishable features of the original signal. And the high pass filter component has little or no resemblance of the original signal. The low pass component is again decomposed into two components. The decomposition process can be continued up to the last possible level or up to a certain desired level. As the high pass filter components have less information discernible to the original signal, we can eliminate the information contents of the high pass filters partially or significantly at each level of decomposition during the reconstruction process. It is this possibility of elimination of the information contents of the high pass filter components that gives higher compression ratio in the case of wavelet based image compression. 3.3 Energy Efficient High Quality Image Transmission scheme In this paper, we propose image transmission scheme driven by energy efficiency considerations in order to be suitable for wireless sensor networks. Wavelet image transform provides data decomposition in multiple levels of resolution, so the image can be divided into packets with different priorities, the packets are ready to be sent. sufficient energy to forward them, if it knows that a node further down the path has an insufficient report the lowest energy level currently available in others nodes. The delay induced by the feedback is proportional to the distance between the concerned nodes. 3.4 Noise Reduction Filter The filters in the Noise Reduction class are designed to remove extreme or outlier values from the transmitted image areas that should have relatively uniform values. These outlier values are often the result of additive noise imposed on the image by the acquisition system or later processing errors. Median, Modal etc., are some of the examples of this type. In this work a study is made on the box filtering techniques used to remove the noise from the transmitted image by EEHQIT Scheme. A spatial averaging filter in which all coefficients are equal is called a box filter. These types of filters are used for blurring and for noise reduction. The output of such a linear smoothing filter is simply the average of the pixels in the neighborhood of the pixel mask. The idea behind smoothing filters is straight forward. By replacing the value of every pixel in an image by the average of the levels in the neighborhood of the filter mask, the process results in an image with reduced sharp transitions. Hence the most obvious application of smoothing is noise reduction. I. EXPERIMENTAL EALUATION ON LOSSLESS IMAGE COMPRESSION TECHNIQUES Fig 1: Image Transmission with Reduced Noise Figure 1 shows the diagrammatic representation of our proposed Image Transmission Approach. The source sensor transmits the packets starting by those with the highest priority, and then continues with those of the next lower priority, and so on. This is carried out using a threshold-based drop scheme where each of the priorities is associated to amount of energy. Of course, the node does not initially know the state-of-charge of the other nodes. This knowledge is gradually obtained from received acknowledgment packets. Thus feedback is used to an energy level. A node can discard packets even if it has The experimental evaluation on image transmission is carried out with JPEG images comparing the noise level with and without using filtering technique and compare energy performance of image transmission schemes in various scenarios. A monochrome image of 18X18 pixels, is used as a test image. This one is 8 bits per pixel originally encoded. That means a data length of bytes, including the image header of 10 bytes. Numerical values adopted for the input parameters of energy models are described below. Then, we present the results of numerical application. To get a reference, we evaluated the consumed energy by transmitting the whole image (3749 bytes) reliably without applying WT or compression algorithms. In the following, we call that the "the original scenario". The amount of energy dissipated to transmit the original image is 15J per hop. Afterwards, we applied WT once and then twice without compression. When WT is applied once, we obtained a resolution 0 of 4106 bytes and a resolution 1 of 188 bytes. Similarly, when WT was applied twice, we obtained 1034, 307 and 188 bytes for resolutions 1, and 3 respectively. We 1 ISSN (PRINT) : , olume -1, Issue -4, 013

4 computed the average energy consumption to transmit the image for scenario (Interactive Image Transmission Energy and Time WT).The parameters which are used in the filter performance evaluation are Signal to Noise Ratio (SNR), Peak Signal to Noise Ratio (PSNR). above approximately 50 db, although this may vary in a minor way for each person Calculation of SNR SNR compares the level of desired signal to the level of background noise. The higher the ratio, the less obtrusive the background noise is. SNR in decibels is defined as SNR 10log Where, e 0 o e is the variance of the noise free image and is the variance of error (between the original and denoised image). Brighter regions have a stronger signal due to more light, resulting in higher overall SNR 38 Calculation of PSNR PSNR is the ratio between possible power of a signal and the power of corrupting noise that affects the fidelity of its representation. PSNR in decibels is easily defined from RMSE as given below PSNR = 0 log10 (55/RMSE) Where, 1 RMSE MN M N y 1 x 1 [ I( x, y) I'( x, y)] where I(x,y) is the original image, I'(x,y) is the decompressed image and M,N represents dimensions of the images. A lower value for MSE means lesser error, and as seen from the inverse relation between the MSE and PSNR, this translates to a high value of PSNR. Logically, a higher value of PSNR is good because it means that the ratio of Signal to Noise is higher. The signal is the original image, and the noise is the error in reconstruction. It is highly required to evaluate a compression scheme having a lower MSE (and a high PSNR).. RESULT AND DISCUSSIONS As a performance measurement, the Peak Signal-to- Noise Ratio (PSNR) is calculated for the reconstructed images at the receiver side. PSNR is metric used to compare two images, the more pixel difference between the images, the less the PSNR value. It is useful to know that the human eye does not have enough sensitivity to detect changes in visual data for PSNR measurements Figure : Imge size vs PSNR Figure depicts the performance graph for showing the result of PSNR value at sender side and receiver side. PSNR value is measured in terms of decibel. At the receiving side, the recovered DC and AC components are used to reconstruct the image. When channel noise is not considered, the image quality keeps improving, in terms of its PSNR value, as we use more AC components. However, to achieve some degree of compression, some of the AC components, those that correspond to the high frequencies in the image, can safely be excluded when reconstructing the image. By using our proposed technique the PSNR value of receiver side is nearly equal to the sender side value. Figure 3: No. Of Images vs PSNR Figure 3 gives the performance graph for showing the result of PSNR value at sender side and receiver side with number of images used. By using our proposed technique the PSNR value of receiver side is nearly equal to the sender side value. 13 ISSN (PRINT) : , olume -1, Issue -4, 013

5 I. CONCLUSION The use of filter in on transmitted compressed digital image improves the image quality to a great extent. Mainly in the case of presence of speckle noise, filtering is very much required in order to improve the diagnostic examination and also to improve the efficiency of post processing techniques like segmentation. Out of the different filters used, Spatial Average filter did the best job as far as synthetic image is concerned. The best result for transmission of compressed image at the receiver end is obtained with good image quality. Proposed filter showed better performance even for the highly dense speckle noise and its efficiency is based on the selection of homogenous region. The homogenous region is selected by the user to arrive the required visual quality for image post processing analysis. The technique adapted in this work selects the best filter threshold automatically for a given image on the basis of statistical parameters and reduces the burden of the user in selecting appropriate filter for different types of images II. REFERENCES [1] Min Wu and Chang Wen Chen. Collaborative image coding and transmission over wireless sensor netowrks. EURASIP Journal on Advances in Signal Processing, 007. Article ID [] G. Zaruba and S. Das, Off-the-shelf enablers of ad hoc networks. New York: IEEE Press Wiley, 003 [3] W. Zhang, Z. Deng, G. Wang, L. Wittenburg, and Z. Xing, Distributed problem solving in sensor networks, in Proceedings of the first international joint conference on Autonomous agents and multiagent systems. ACM Press, 00, pp [4] Liu, C.-M., Lee, C.-H., and Wang, L.-C.. Powere±cient communication algorithms for wireless mobile sensor networks. In 1st ACM International Workshop on Performance Evaluation of Wireless, Ad Hoc, Sensor and Ubiquitous Networks, 004 pages [5] Magli, Mancin, M., and Merello, L.. Low complex- ity video compression for wireless sensor networks. In Proceedings of 003 International Conference on Multi- media and Expo, 003 pages [6] W. Yu, Z. Sahinoglu, and A. etro, Energy efficient JPEG 000 image transmission over wireless sensor networks, in Proceedings of IEEE Global Telecommunications Conference (GLOBECOM 04), vol. 5, pp , Dallas, Tex,USA,November- December 004. [7] W.Wang,D. Peng, H.Wang,H. Sharif, andh.h. Chen, Optimal image component transmissions inmultirate wireless sensor networks, in Proceedings of the 50th Annual IEEE Global Communications Conference (GLOBECOM 07), Washington, DC, USA, November 007. [8] C. Schurgers, O. Aberthorne, and M. B. Srivastava, Modulation scaling for energy aware communication systems, in Proceedings of the International Symposium on Low Power Electronics and Design (ISLPED 01), pp , Huntington Beach, Calif, USA, August 001. [9] A. Ephremides, Energy concerns in wireless networks, IEEE Wireless Communications, vol. 9, no. 4, pp , August 00. [10] Alla Aksel, Andrew D. Gilliam, John A. Hossack, Scott T. Acton, SPECKLE REDUCING ANISOTROPIC DIFFUSION FOR ECHOCARDIOGRAPHY, ACSSC 40th Asilomar Conference on Signals, Systems and Computers, 007,pages: [11] Yang Mo Yoo, Fan Zhang, Liang Mong Koh, and Yongmin Kim, Nonlinear Diffusion in Laplacian Pyramid Domain for Ultrasonic Speckle Reduction, IEEE TRANSACTIONS ON MEDICAL IMAGING, 007, OL.6, pages: [1] Ricardo G. Dantas and Eduardo T. Costa, Ultrasound Speckle Reduction Using Modified Gabor Filters, IEEE TRANSACTIONS ON ULTRASONICS, FERROELECTRICS AND FREQUENCY CONTROL, 007, OL.54, pages: [13] F. Labeau, J. C. Chiang, M. Kieffer, P. Duhamel, L. andendorpe, and B. Marq. Oversampled filter banks as error correction codes: Theory and impulse noise correction. submitted to IEEE Trans. on Signal Processing, 004. [14] G. Rath and C. Guillemot. Characterization of a class of error-correcting frames and their application to image transmission. In Proceedings of PCS, St Malo, ISSN (PRINT) : , olume -1, Issue -4, 013