Mohammed E. Al-Mualla

Dual-tree complex wavelet coefficient magnitude modelling using the bivariate Cauchy-Rayleigh distribution for image denoising

Abstract Wavelet shrinkage is a standard technique for denoising natural images. Originally proposed for univariate shrinkage within the Discrete Wavelet Transform (DWT) domain it has been made more effective through the use of (approximately) translationally invariant wavelet decompositions such as the Dual-Tree Complex Wavelet Transform (DT-CWT) and bivariate shrinkage. Techniques using the DT-CWT benefit from both (approximate) translational invariance and improved directionality. However, these techniques have used the real and imaginary components of the complex valued subband coefficients separately. The proposed work instead uses coefficient magnitudes to produce a novel bivariate shrinkage technique based on a heavy tailed bivariate distribution (of magnitudes) to provide a quantitative improvement in image denoising. The results were compared to state of the art non-local means denoising technique BM3D. The PSNR results for small amounts of noise were comparable and within a small range for larger amounts of noise. However, when using the perceptually based structural similarity metric (SSIM) our developed technique offers improved results across the range of noise inputs when compared to BM3D in many cases. Perceptually, the developed technique is able to retain a greater quantity of the high frequency elements of the input image compared to BM3D.

Undecimated Dual-Tree Complex Wavelet Transforms

Two undecimated forms of the Dual Tree Complex Wavelet Transform (DT-CWT) are introduced together with their application to image denoising and robust feature extraction. These undecimated transforms extend the DT-CWT through the removal of downsampling of filter outputs together with upsampling of the complex filter pairs in a similar structure to the Undecimated Discrete Wavelet Transform (UDWT).Both developed transforms offer exact translational invariance, improved scale-to-scale coefficient correlation together with the directional selectivity of the DT-CWT. Additionally, within each developed transform, the subbands are of a consistent size. They therefore benefit from a direct one-to-one relationship between co-located coefficients at all scales and therefore this offers consistent phase relationships across scales. These advantages can be exploited within applications such as denoising, image fusion, segmentation and robust feature extraction. The results of two example applications (bivariate shrinkage denoising and robust feature extraction) demonstrate objective and subjective improvements over the DT-CWT. The two novel transforms together with the DT-CWT offer a trade-off between denoising performance, computational efficiency and memory requirements. HighlightsProposed transforms have exact translational invariance.Coefficients have one-to-one cross scale relationships.Improved results for two example applications.Matlab code available at: www.bristol.ac.uk/vi-lab/projects/udtcwt.

Autonomous detection and tracking under illumination changes, occlusions and moving camera

In this paper, an autonomous multiple target detection and tracking technique for dynamic scenes that are influenced by illumination variations, occlusions and camera instability is proposed. The framework combines a novel Dynamic Reverse Analysis (DRA) approach with an Enhanced Rao-Blackwellized Particle Filter (E-RBPF) for multiple target detection and tracking, respectively. The DRA method, in addition to providing accurate target localization, presents the E-RBPF scheme with costs associated with the differences in intensity caused by illumination variations between consecutive frame pairs in any video of a dynamic scene. The E-RBPF inherently models these costs, thus allowing the framework to (a) adapt learning parameters, (b) distinguish between camera-motion and object-motion, (c) deal with sample degeneracy, (d) provide appropriate appearance compensation during likelihood measurement and (e) handle occlusion. The proposed detect-and-track method when compared against other competing baseline techniques has demonstrated superior performance both in accuracy and robustness on challenging videos from publicly available datasets. HighlightsAnalysis of illumination changes in a scene via forward & reverse background modeling..Generating sufficient statistics to characterize variations in illumination.Building a hybrid model of the background using selected frames.Optimized implementation of reverse analysis for trade-off between accuracy and load.TTightly integrated likelihood and noise models for robust Rao-Blackwellised Particle filtering.

2-D Walsh coding for robust digital image watermarking

One-dimensional (1D) Walsh coding have previously been proposed to improve the robustness of digital image watermarking techniques. In this paper, it is proposed to use two-dimensional (2D) Walsh coding to achieve further improvements in robustness. In the proposed method, Walsh functions are used to perform 2D coding of the watermark before embedding it in the original image. Simulation results, using both spatial-domain and frequency-domain techniques, indicate that the use of 2D Walsh coding results in improved robustness against JPEG compression attacks compared to 1D Walsh coding. Further improvements can also be achieved by increasing the length of the Walsh sequence. This is, however, at the expense of more degradation to the watermarked image.

Walsh-coded signatures for robust digital image watermarking

This paper proposes the use of 1-D Walsh coding to improve the robustness of digital image watermarking techniques. In the proposed method, Walsh functions are used to encode the watermark before embedding it in the original image. Results indicate that this type of coding results in less visible degradation to the watermarked image than direct scaling. Simulation results using both spatial-domain and frequency-domain techniques indicate that the use of Walsh coding results in improved robustness against JPEG compression attacks. It is also shown that frequency domain watermarking is more resistant to JPEG compression attacks and also less sensitive to the contents of the input original image. Increasing the length of the Walsh function results in more robustness but at the expense of more degradation to the quality of the watermarked image.

Perceptual Image Fusion Using Wavelets

A perceptual image fusion method is proposed that employs explicit luminance and contrast masking models. These models are combined to give the perceptual importance of each coefficient produced by the dual-tree complex wavelet transform of each input image. This combined model of perceptual importance is used to select which coefficients are retained and furthermore to determine how to present the retained information in the most effective way. This paper is the first to give a principled approach to image fusion from a perceptual perspective. Furthermore, the proposed method is shown to give improved quantitative and qualitative results compared with previously developed methods.

Low Complexity Power Optimization Algorithm for Multimedia Transmission Over Wireless Networks

This paper considers the problem of transmit power optimization for multimedia applications in continuous high-speed transmission over wireless networks. The power optimization process is developed by noting that some performance metrics such as throughput, delay and peak signal-to-noise ratio (PSNR) for particular systems with hybrid automatic repeat request (HARQ) may exhibit a staircase behavior. In such scenarios, the corresponding metric remains fixed for a wide range of signal-to-noise ratios (SNRs). Consequently, the transmit power can be reduced significantly while the relevant metric remains almost unchanged. The obtained results reveal that invoking power optimization algorithms can achieve a significant power saving of about 80% for particular scenarios. The system considered in this work is a truncated HARQ with turbo product codes (TPC) and parallel concatenated convolutional codes (PCCC). Chase combining is also used to combine the retransmitted packets with the original transmission. An efficient semi-analytical model is developed to obtain the system throughput in additive white Gaussian noise (AWGN) and Rayleigh fading channels. The obtained results also show that using the throughput, delay or PSNR as performance metrics provides equivalent power saving results.

Adaptive Hybrid ARQ System Using Turbo Product Codes with Hard/Soft Decoding

The bit error rate (BER) performance of turbo product codes (TPC) has been considered extensively in the literature. However, other performance metrics such as throughput can be more informative in particular systems. In this letter, the throughput performance of hybrid automatic repeat request (HARQ) is considered using TPC with iterative hard and soft decision decoding. Monte Carlo simulation and semi-analytical solutions are developed to evaluate the throughput of HARQ-TPC system for a wide selection of codes. The obtained results reveal that the coding gain advantage of the soft over hard decoding is reduced significantly when throughput is adopted as the performance metric, and it actually vanishes completely for some codes. When adaptive coding is used, the soft decoding advantage is limited to about 1.4 dB.

Supervised remote sensing image segmentation using boosted convolutional neural networks

In this paper, a region segmentation technique for remote sensing images using a boosted committee of Convolutional Neural Networks (CNNs) coupled with inter-band and intra-band fusion, is proposed. The vast heterogeneity in remote sensing images restricts the application of existing segmentation methods that often rely on a set of predefined feature detectors along with tunable parameters. Therefore, it is highly challenging to design a segmentation technique which could achieve high accuracy while simultaneously maintaining strong generalization particularly for visual data with improved spatial, spectral, and temporal resolutions. The proposed method is a fusion framework consisting of a set of thirty boosted networks that derive individual probability maps on the location of region boundaries from the different multi-spectral bands and combines them into one using an averaging inter-band fusion scheme. The boundaries are then thinned, connected, and region segmented using a morphological intra-band fusion scheme. Qualitative and quantitative results, on publicly-available datasets, confirm the superiority of the proposed segmentation method over existing state-of-art techniques. In addition, the paper also demonstrates the effect of some variations in design-choices of the proposed method.

Performance evaluation of direction finding algorithms for adapative antenna arrays

Adaptive antenna arrays have been widely adopted in third-generation (3G) mobile systems due to their capability of increasing the coverage range and capacity of base stations by directing beam patterns towards the desired signals and null-patterns towards the interferers. A key to such capability is the use of Direction-of-Arrival (DOA) techniques. This paper evaluates and compares the performance of three high-resolution DOA algorithms. In particular, the paper investigates the boundaries within which the performance of the algorithms is considered acceptable.