Performance Evaluation of JPEG Image Transmission over Wireless Fading Channels with Diversity Receivers

Performance Evaluation of JPEG Image Transmission over Wireless Fading Channels with Diversity Receivers

Priyanka Hazowary

ECE, North Eastern Hill University, Shillong,793022, Meghalaya, India

Sushanta Kabir Dutta

ECE, North Eastern Hill University, Shillong,793022, Meghalaya, India

Rahul Raj

ECE, North Eastern Hill University, Shillong,793022, Meghalaya, India

Rupaban Subadar

ECE, North Eastern Hill University, Shillong,793022, Meghalaya, India

Abstract – Wireless transmission of JPEG compressed images is highly susceptible to channel impairments such as noise, fading, and packet loss, which significantly degrade visual quality. This paper presents an enhanced JPEG image transmission framework over wireless channels, focusing on performance improvement through the use of diversity receiver techniques. Image quality at the receiver is evaluated under different modulation schemes and channel conditions using objective metrics such as Peak Signal-to-Noise Ratio (PSNR), Mean Squared Error (MSE), and Bit Error Rate (BER). In addition, visual inspection is employed to assess perceptual image quality. Simulation results demonstrate that diversity receivers significantly improve the reliability and quality of JPEG image transmission, achieving higher PSNR and lower distortion compared to conventional single-receiver systems. The results confirm that the proposed approach provides an efficient and robust solution for high-quality image delivery in wireless communication environments.

Keywords – Diversity combining, image transmission, wireless communications, JPEG, DCT

1. INTRODUCTION

   With the rapid growth of multimedia communication over wireless networks, the transmission of data, voice, images, and video has become increasingly significant. The widespread adoption of portable communication devices has further intensified the demand for efficient and reliable wireless image transmission. However, wireless channels are inherently susceptible to fading and interference, resulting in burst errors and performance degradation that are far more severe than in stable wired environments. To achieve reliable image delivery under such adverse conditions, robust source and channel encoding techniques are required to effectively combat channel impairments. Moreover, the limited bandwidth available in wireless systems makes efficient image compression indispensable for practical and high-quality multimedia communication. Wireless communication channels are affected by several impairments, including interference, delay spread, attenuation, noise, and fading [12]. Among these, multipath fading and noise constitute the most critical sources of signal degradation. These effects arise because transmitted signals propagate along multiple paths due to reflection, diffraction, and scattering in the propagation environment. At the receiver,

   the superposition of these multipath components may result in either constructive or destructive interference depending on their relative phases, which can lead to partial or complete signal cancellation [35]. Such distortions introduce detection errors and significantly degrade overall system performance [6]. To enhance transmission efficiency and reliability, source coding techniques are employed to eliminate redundancy in the data, thereby reducing the amount of information to be transmitted. This compression process plays a vital role in the efficient transmission of various types of data over bandwidth- limited wireless channels. Irgan et al. [7] introduced a cost- effective image prioritization technique in which important image regions are identified and transmitted through more reliable communication paths. Shyam Lal and Mahesh Chandra

   [8] employed wavelet transformbased methods along with a hybrid filtering approach for image denoising in the presence of additive white Gaussian noise (AWGN). The proposed hybrid filter combines a fourth-order partial differential filter with a nonlinear bivariate shrinkage function. Although high peak signal-to-noise ratio and low mean square error are commonly used performance measures, these metrics do not necessarily reflect perceptual image quality; therefore, a feature similarity index was adopted for image quality assessment. Lecuire et al. [9] investigated packet prioritization using the Discrete Wavelet Transform (DWT), in which sensor nodes schedule data transmission according to the correlation between residual battery energy and packet priority. Error correction and detection techniques [1014], such as automatic repeat request (ARQ) schemes and various channel coding strategies, are widely employed to counteract errors introduced during wireless transmission. In addition, error-resilient image coding methods [1520] are designed to reduce the sensitivity of the transmitted bit stream to channel impairments, thereby facilitating more effective error recovery at the receiver. As a result, transmission errors in such coding schemes have a less detrimental impact on reconstructed image quality compared to conventional image coding methods such as JPEG. To further mitigate the adverse effects of fading, diversity techniques are commonly adopted, among which selection combining (SC) is one of the most popular due to its low implementation complexity. Although considerable research has been devoted to analysing the performance of diversity receivers under

   various fading conditions and to improving image compression efficiency and visual quality, relatively few studies have addressed the transmission of compressed images over wireless channels along with a systematic evaluation of post-reception image quality. This work seeks to address this gap by investigating the influence of image compression on wireless image transmission and by assessing the effectiveness of diversity receivers in preserving reconstructed image quality.

   The remainder of this paper is organized as follows: Section II discusses the channel model and compression techniques. Section III presents the modified image compression methods. The mathematical analysis of diversity receivers is provided in Section IV. Section V presents the simulation results and their analysis, followed by the conclusion in Section VI.

2. THE COMPRESSION TECHNIQUES AND CHANNEL

   A block diagram illustrating the complete work is presented in Figure1. A standard image is compressed using a modified algorithm and then transmitted over wireless channels under different modulation schemes. To further enhance image quality, selection diversity is applied at the receiver. The image is restored, and various performance parameters are evaluated. The following sections provide a detailed explanation of each block.

   Fig. 1. A block diagram representation of our work

   1. JPEG

      Joint Photographic Experts Group is an international standard for still picture compression that employs the 2-D Discrete Cosine Transform (DCT), which converts a spatial representation to a frequency representation. Since lower frequencies are more noticeable in an image than higher frequencies, we can reduce the amount of information required to represent the image without significantly compromising image quality if we break an image down into its frequency components and discard many higher frequency coefficients [21].

      In order to reduce or eliminate information that is not as significant to the human eye, the JPEG compression algorithm analyses the image. This includes both high- frequency details that are too small to be seen and some elements like colur variations that are too delicate to be

      recognized. As a result, files are smaller and easier to store and send across networks. A compressed JPEG image preserves perceptually significant visual information while drastically reducing file size. However, blocking and ringing are examples of artefacts that might result from aggressive compression.

      Visualizing the JPEG Pipeline may be as given below:

      Fig. 2. A block diagram representation of JPEG compression

      Despite being widely used JPEG has a few limitations like reduced image quality due to lossy compression, artifacts like blurring and ringing may occur at higher compression ratios, not suitable for sharp edge graphics like line drawings, texts etc.

   2. JPEG 2000

      The Joint Photographic Experts Group (JPEG) created JPEG 2000, a sophisticated picture compression standard, in the early 2000s. It is an advancement over the original JPEG format and has a number of advantages over other image compression formats, including enhanced image quality, increased compression efficiency, and support for sophisticated features like region of interest encoding and scaling. Because the JPEG 2000 compression algorithm is based on wavelet transforms, it can compress a picture more effectively and adaptably than other compression techniques. The technique creates tiles, which are tiny portions of the image that can be separately compressed.

   3. Modified JPEG with DCT

      The frequency domain image transform technique known as DCT is used to minimize the amount of storage space required for the image. The entire image is first split up into n×n blocks, which are then subjected to DCT. High order words are used to indicate changes in the blocks’ width and height, while low order terms are used to express the average value inside a block [22]. The image can be restored from compressed format using the IDCT (Inverse Discrete Cosine Transform). In lossy compression methods, DCT is especially utilized.

      The 2-D DCT is an expansion of the 1-D case:

      complex exponentials, it solely employs real-valued functions. The DST analyses periodic signals with odd

      symmetric at a midway and begin and stop at zero. The DST is particularly effective in image and audio compression applications, as it may represent a signal with fewer coefficients than the DFT.

      Image Quality and Less visual optimization: Since all frequencies are quantized equally, low-frequency details (which contribute most to visual quality) are compressed just as aggressively as high-frequency details. As a result, image quality deteriorates further with more noticeable loss of details and artifacts being more pronounced. However, it gives us a superior noise reduction and even distortion across the image and therefore securing targeted data loss. This may be particularly useful for experimental image processing like transmission in wireless channels.

   4. Modified JPEG with DST

   The Discrete Sine Transform (DST) is a mathematical technique employed in signal processing and image compression to transform a sequence of discrete data samples into a set of sine function components. It is comparable to the Discrete Fourier (DFT), except instead of

   2N

   where s(u, v) represents frequency response value for u , v & f (x, y) represents pixel color value at position (x, y) [23].

   In certain situations, DST provides better energy compaction than DCT when used with modified JPEG, making it more useful for specific image patterns. Additionally, in certain cases, it might offer marginally better visual quality or compression efficiency than DCT. In contrast, it can be said that DCT is faster, more effective, and more appropriate for images that are smooth and realistic, but it is also prone to blocking artefacts. Conversely, DST is helpful for texts with more computational complexity or visuals with sharp edges, such as line drawings.

3. WIRELESS CHANNEL

   Wireless transmission channels are widely recognized to be affected by burst-type errors. These errors primarily arise due to physical phenomena such as fading and multipath propagation. Accurately modeling the physical characteristics of a wireless channel is a challenging task, as it depends on factors including the motion of the transmitter, receiver, and surrounding objects along the signal path [25]. Although various models have been introduced in the literature to describe these physical effects, this work adopts an input output channel model, which focuses on representing the relationship between the transmitted and received signals.

   In recent years, wireless communication has attracted significant attention. Due to multipath propagation, the received signal consists of multiple attenuated, delayed, and phase-shifted copies of the original transmitted signal. When JPEG images are transmitted over noisy wireless channels, the presence of such impairments can severely degrade the quality of the reconstructed visual information. Moreover, the use of

   Variable Length Coding (VLC) increases sensitivity to transmission errors, as even minor bit errors may propagate and cause noticeable distortion. Consequently, designing an

   input branches. The CDF of single branch can be calculated from (5) by performing the integration, which is given below [26]

   efficient and reliable system for image transmission over

   th 1

   wireless channels remains a challenging task. Numerous approaches have been proposed to address this issue [24].

   1. Rayleigh fading

      The Rayleigh distribution is commonly employed to characterize multipath fading in scenarios where a large number of independently scattered signal components and no direct line-of-sight (LOS) path is received by the receiver. Under these conditions, the probability density function (PDF)

      of the channel fading amplitude can be expressed as follows

      If we consider L number of branches, in the SC receiver the CDF of the output SNR can be obtained by taking product of L terms of equation (6).

       

      where X and Y are independent Gaussian RVs with zero mean and equal variances.

   2. Diversity Receivers

   Diversity is a widely used method to eliminate the effect of small-scale fading. In diversity receivers signals are transmitted through multiple independent fading paths to

   The above expression can be used to derive the various performance measures of the communication system.

   The error probability for various modulation schemes over Rayleigh fading channel considering SC receiver can be obtained by averaging the conditional probability of error ( pe (| ) ) over the PDF given in (8).

   sc

   reduce the probability of deep fading. The multiple fading paths can be achieved through various methods, such as time

   diversity, frequency diversity, space diversity etc. Among these methods, space diversity used multiple antennas to

   pe

   0

   pe (| sc ) f ( sc )d sc

   (9)

   receive the transmitted signals. These antennas are placed few (wavelength) apart from each other to ensure independent fading path. Various diversity receiver techniques combine this received signal to enhance the SNR of the received signal

   For binary coherent modulation pe (| sc ) Q 2a sc ,

   where a=0.5 for binary FSK and a=1 for BPSK. Now putting the values in (9) we obtained

   in different way. Out of different diversity techniques, SC is

   L

   L1

   the simplest one from implementation point of view. In SC

   pe Q 2a sc 1 exp sc

   receiver the signals received by the various antennas are

   compared continuously and signal with highest SNR is decoded

    

4. ATHEMATICAL ANALYSIS OF THE MODEL

   After The SNR pdf of Rayleigh fading channel can be derived from (3) by performing the random variable

   sc

    

   The above integral is solved by using the binomial theorem and using the equations [27(A-8a and A-6)]

   transformation according to the following formula

   where Eb is the energy per bit and

   N0 is the power spectral

   2 F1 1,1.5, 2, k 1 a

   density of Gaussian noise.

   The final result available in [10]

   In a similar way the BER of SC receiver can be evaluated for various modulation schemes.

   For SC receiver the output SNR pdf can be derived by taking the derivative of product CDF of all the

   The peak signal to noise ratio of an image is mathematically evaluated using the formula given below

   2552

   (c) Lena (d) Mandrill.

   PSNR 10 log10 MSE (dB) (12)

   Where, MSE

   M N is the size of the image matrix, f ‘ (i, j) is the

   received image and

   f (i, j) is the transmitted image.

5. RESULTS AND DISCUSSIONS

   The proposed scheme is applied to all test images. Consequently, results are obtained using the proposed transforms with diversity and without diversity. We have plotted the Peak-SNR (PSNR) vs Average SNR per branch for images of Cameraman, Peppers, Lena, and Mandrill respectively for the tabulated results in Fig. 4 through Fig. 7. These are performed to compare the received image quality after applying varying amounts of compression to the original images used for transmission in BPSK and QPSK. The Peak Signal-to-Noise Ratio (PSNR) is the most widely used metric for evaluating the quality of reconstructed or received images. The mathematical equations for MSE and PSNR were previously described. A lower Mean Square Error (MSE) value signifies fewer errors, reflecting the inverse relationship between MSE and PSNR. As the MSE decreases, the PSNR correspondingly increases. A higher PSNR indicates that the signal strength is much greater relative to the noise ratio. The received image quality improves as the PSNR value increases.

   In Fig. 8 through Fig. 11 show the bar graphs of PSNR vs Average SNR per branch at different compression ratios for Cameraman, Peppers, Lena, and Mandrill respectively. Diversity techniques significantly enhance image quality, thereby, making wireless image transmission more robust against fading and noise.

   Fig. 4. The curves of Peak SNR vs Average SNR per branch for different quality compression for Cameraman image.

   Fig. 5. The curves of Peak SNR vs Average SNR per branch for different quality compression for Peppers image.

   Fig. 6. The curves of Peak SNR vs Average SNR per branch for different quality compression for Lena image.

   Fig. 3. Original image of (a) Cameraman (b) Peppers

   Fig. 7. The curves of Peak SNR vs Average SNR per branch for different quality compression for Mandrill image.

   Fig. 11. Bar Graph of Peak SNR vs Average SNR per branch for different quality compression for Mandrill image.

   Fig. 8. Bar Graph of Peak SNR vs Average SNR per branch for different quality compression for Cameraman image.

   Fig. 12. Bar Graph of Peak SNR vs Average SNR per branch for different quality compression for Mandrill.

   Table I. COMPRESSION PERFORMANCE OF DIFFERENT IMAGE QUALITY IN PSNR AND MSE FOR CAMERAMAN IMAGE

   SNR	PSNR without diversity	PSNR with diversity
   0	8.4188	10.8594
   2	9.6126	13.1776
   4	11.1534	15.7568
   6	12.7841	18.5792
   8	14.4082	22.0582
   10	16.2559	25.0473
   12	18.2837	29.4725
   14	20.0894	33.5408
   16	22.2765	36.7035
   18	24.4173	40.3833
   20	25.8348	48.1648

    

   Fig. 9. Bar Graph of Peak SNR vs Average SNR per branch for different quality compression for Peppers image.

   Fig. 10. Bar Graph of Peak SNR vs Average SNR per branch for different quality compression for Lena image.

   Table II. COMPRESSION PERFORMANCE OF DIFFERENT IMAGE QUALITY IN PSNR AND MSE FOR PEPPERS IMAGE

   <td21.8504

   SNR	PSNR without diversity	PSNR with diversity
   0	8.3462	10.8562
   2	9.7225	13.0701
   4	11.0679	15.5832
   6	12.7266	18.4753
   8	14.4063	21.6617
   10	16.3492	24.8810
   12	18.1648	28.3876
   14	20.1921	33.5408
   16	37.3730
   18	24.3087	42.1442
   20	25.8859	45.1545

   Table III. COMPRESSION PERFORMANCE OF DIFFERENT IMAGE QUALITY IN PSNR AND MSE FOR LENA IMAGE

   SNR	PSNR without diversity	PSNR with diversity
   0	8.3876	10.9848
   2	9.6626	12.9797
   4	11.0721	15.6484
   6	12.7866	18.4706
   8	14.3195	21.8100
   10	16.3521	25.5880
   12	18.2304	29.1885
   14	20.2619	32.4828
   16	22.0689	38.1648
   18	24.3446	48.1648
   20	26.5511	48.1648

   Table IV. COMPRESSION PERFORMANCE OF DIFFERENT IMAGE QUALITY IN PSNR AND MSE FOR MANDRILL IMAGE

   SNR	PSNR without diversity	PSNR with diversity
   0	8.3469	10.9141
   2	9.6620	13.0473
   4	11.1649	15.5466
   6	12.7836	18.5253
   8	14.4813	21.7533
   10	16.3511	25.2274
   12	18.2436	28.6546
   14	20.2036	33.3455
   16	21.9808	36.2150
   18	24.0598	41.7965
   20	26.1782	45.1544

   The difference in PSNR values between systems with and without diversity increases as the signal-to-noise ratio (SNR) rises, indicating that diversity techniques exert a stronger

   influence on signal quality under favorable channel conditions. The gray bars illustrate that diversity-based reception yields the highest PSNR values, particularly at higher average SNR levels. In contrast, the orange bars, representing transmission without diversity, exhibit only moderate improvement over the baseline and demonstrate limited robustness. This behavior suggests that non-diversity systems are more vulnerable to channel impairments and less reliable under challenging propagation conditions. Overall, the application of diversity techniques substantially enhances the reliability and robustness of wireless communication systems, especially in environments affected by fading and interference. Notably, the improvement in PSNR becomes more pronounced at higher data indices, confirming the effectiveness of diversity in preserving image quality. Table I IV present the experimental PSNR results obtained using the Modified DCT-JPEG scheme at different SNR levels for the Cameraman, Peppers, Lena, and Mandrill images, respectively. Each table compares the performance of the system over a Rayleigh fading channel with and without the application of diversity techniques, for SNR values ranging from 0 dB to 20 dB in steps of 2dB. The results clearly indicate that the incorporation of diversity leads to a consistent improvement in overall system performance. Specifically, the average reconstruction quality increases to approximately 48.1648 dB, while the corresponding error level is reduced to an average of 25.5880dB. Furthermore, the experimental results indicate that the proposed algorithm outperforms conventional JPEG across the entire SNR range, achieving superior PSNR values as well as enhanced visual quality. At low SNR values, especially around 0 dB, the reconstructed images show significant degradation in quality and correspond to the highest compression ratios. In contrast, at higher SNR levels such as 20 dB, the reconstruction quality improves substantially, and lower compression ratios are achieved. These results indicate that both image quality and compression ratio can be adaptively adjusted according to channel conditions and application requirements. The advantage of the proposed compression algorithm is the uniform quantization used for transform coefficients, especially saving much multiplication operations when adjusting the bit rates. In Fig. 12 we have plotted the BER Vs Average SNR per branch curve for various diversity order. As expected, the BER performance improve with increase in order of diversity.

6. CONCLUSIONS

This study evaluated a modified image compression technique designed for wireless transmission with diversity, balancing compression efficiency and image quality. The results demonstrate that the proposed method effectively enhances image fidelity while ensuring reliable transmission across varying wireless channel conditions. By leveraging diversity schemes, the approach mitigates signal degradation, reducing transmission errors and improving overall visual quality. Performance metrics such as PSNR MSE, and average SNR confirm the robustness of the technique, making it a promising solution for wireless multimedia applications.

REFERENCES

1. Sherwood, P. G., & Zeger, K., Error protection for progressive image transmission over memoryless and fading channels. IEEE Transactions on Communications, 46(12), 1555-1559, (1998).
2. Chande, V., & Farvardin, N., Progressive transmission of images over memoryless noisy channels. IEEE Journal on Selected Areas in Communications, 18(6), 850-860.
3. Biglieri, E., Proakis, J., & Shamai, S., Fading channels: Information- theoretic and communications aspects. IEEE transactions on information theory, 44(6), 2619-2692 (2000).
4. Sherwood, P. G., & Zeger, K., Progressive image coding on noisy channels. In Proceedings DCC’97. Data Compression Conference (pp. 72-81). IEEE (1997, March).
5. Boulgouris, N. V., Thomos, N., & Strintzis, M. G., Transmission of images over noisy channels using error-resilient wavelet coding and forward error correction. IEEE Transactions on Circuits and Systems for Video Technology 13.12 (2004): 1170-1181.
6. Baharuddin, Performance Analysis of SPIHT Compressed Image Transmission with Diversity Selection Combining on Radio Frequency 10th International Conference on Quality in Research (QIR) vol 1 pp ICT-12 (1-4) (2007)
7. K. Irgan, C. Ünsalan, and S. Baydere, Low-cost prioritization of image blocks in wireless sensor networks for border surveillance, J. Netw.Comput. Appl., vol. 38, pp. 5464, Feb. 2014.
8. Shyam LaI and Mahesh Chandra (2012) “Image Denoising of Wavelet based Compressed Images Corrupted by Additive White Gaussian Noise” Research Journal of Applied Sciences,Engineering and Technology 4 (17):3108-311 8© Maxwell Scientific Organization.
9. Lecuire, V., Duran-Faundez, C., & Krommenacker, N., Energy- efficient transmission of wavelet-based images in wireless sensor networks. Eurasip journal on image and video processing, 2007, 1-11.
10. Makhoul, J., A fast cosine transform in one and two dimensions. IEEE Transactions on Acoustics, Speech, and Signal Processing 28 (1): 27 34.doi:10.1109/TASSP.1980.1163351 (February 1980).
11. Sklar, ., Digital Communications: Fundamentals and Applications,

    Second Edition, Prentice-Hall, 2001, pp 436-460.

12. Y. H. Han and J. J. Leou, Detection and correction of transmission errors in JPEG images, IEEE Trans. Circuits Syst. Video Technol., vol. 8, pp. 221231, Apr. 1998.
13. Y. Koyama and S. Yoshida, Error control for still image transmission over a fading channel, in Proc. IEEE 45th Veh. Technol. Conf., vol. 2,1995, pp. 609613.
14. I. Kozintsev, J. Chou, and K. Ramchandran, Image transmission using arithmetic coding-based error detection, in Proc. IEEE Data Compression Conf., 1998, pp. 339348.
15. X. Fei and T. Ko, Turbo-codes for compressed image transmission over frequency selective fading channel, in Proc. IEEE 6th Int. Conf. Universal Personal Commun., vol. 2, 1997, pp. 505509.
16. R. E. Van Dyck and G. M. Horiban, Wavelet video transmission over tactical wireless channels, in Proc. 1st Int. Conf. Wireless Image/Video Commun., Sept. 1996, pp. 7479.
17. Z. Sun, J. Luo, C. W. Chen, and K. J. Parker, Analysis of a wavelet- based compression scheme for wireless image communication, Proc.SPIE, vol. 2762, pp. 454465, 1996.
18. E. S. Jang and N. M. Nasrabadi, Subband coding with multistage VQ for wireless image communication, IEEE Trans. Circuits Syst. VideoTechnol., vol. 5, pp. 247253, June 1995.
19. T. H. Meng, B. M. Gordon, E. K. Tsern, and A. C. Hung, Portable video-on-demand in wireless communication, Proc. IEEE, vol. 83, pp.659680, Apr. 1995.
20. S. S. Hemami, Reconstruction-optimized lapped orthogonal transforms for robust image transmission, IEEE Trans. Circuits Syst. VideoTechnol., vol. 6, pp. 168181, Apr. 1996.
21. Falila, S. R., et al., A comparative study of different multi-spectral images compression methods. 2013 Second International Japan-Egypt Conference on Electronics, Communications and Computers (JEC- ECC). IEEE, 2013.
22. Rani, P., & Arora, A, A survey on comparison between DCT and DWT techniques of image compression. International Journal of Science and Research (IJSR) (2015): 2319-7064.
23. Hazowary, P., Talukdar, N., Dutta, S.K., Subadar, R, A Comparative Study of Jpeg and Similar Compression Standards for Wireless Communication. NIELIT’s International Conference on Communication, Electronics and Digital Technologies. Singapore: Springer Nature Singapore, 2024.
24. Over Wireless Channel. International Journal of Scientific and Research Publications 2.5 (2012): 1-5.
25. L. N. Kanal and A. R. K. Sastry, Models for channels with memory and their applications to error control, Proc. IEEE, vol. 66, pp. 724744,

    July 1978.

26. T. S. Rappaport, Wireless communication principles and practice,

    New Jersey: Prentice Hall, 1996.

27. Aalo, V. A. (2002). Performance of maximal-ratio diversity systems in a correlated Nakagami-fading environment. IEEE transactions on Communications, 43(8), 2360-2369.