Expanding window fountain (EWF) codes use a windowing technique to achieve unequal error protection (UEP) and unequal recovery time. UEP affords significant performance improvement of the more important bits (MIB), but at the cost of performance deterioration of the less important bits (LIB), resulting in poor intermediate performance of the EWF coding. In this study, the authors propose the utilisation of intermediate feedback in EWF coding to terminate the encoding of the MIB after their recovery, with the purpose of enhancing the intermediate performance of EWF codes over binary input additive white Gaussian noise (BIAWGN) channels. The effectiveness of the proposed coding scheme is analysed and evaluated by using extrinsic information transfer functions to investigate the theoretical performance and obtain the closed‐form formula of the bit error rate (BER) of the LIB after the intermediate feedback. The results of the theoretical analysis and simulation of the scheme using additive white Gaussian noises with different signal‐to‐noise ratios reveal an improvement in the BER performance of the LIB compared to the case of conventional EWF coding, and also show that the scheme significantly enhances the intermediate performance of the EWF codes and improves the transmission efficiency.

In this paper, we investigate feedback in EWF codes. The considered type of feedback is an acknowledgment, which is a single bit per source block to inform the most important bits have been decoded completely. Moreover, we propose a practical prediction model to calculate the average overheads of successful decoding for different importance classes of the source block. Asymptotic analysis and simulation results demonstrate that, compared with standard EWF codes, the proposed EWF codes with feedback can obtain a lower average overhead of successful decoding in lower prioritized data, and the proposed prediction model is effective.

Novel precoded Expanding Window Fountain (EWF) codes, based on low rate low-density parity-check (LDPC) codes, are investigated for finite message length over binary erasure channels (BECs). The proposed codes are obtained by utilizing low rate LDPC codes as precodes and degree distributions with high intermediate symbol recovery rate for EWF codes. Simulation results show that, compared with conventional precoded EWF codes, our proposed codes have better performance in the scenarios of small and moderate message lengths over the BEC.

In this study, the authors propose an improved design of expanding window fountain (EWF) codes over the binary input additive white Gaussian noise channels. The encoding process is modified to get almost regular variable‐node degree distributions. Compared with the conventional EWF codes, their proposed scheme can lower the error floor. Furthermore, they combine the EWF encoding process with unequal power allocation to improve unequal error protection (UEP) performance, and construct pre‐coded EWF codes by concatenating their proposed EWF codes with high‐rate low‐density parity‐check codes. Both asymptotic analysis and simulation results show that, compared with the conventional scheme, their proposed scheme can get better UEP performance and design flexibility.

In this paper, we propose an efficient scheme to construct fountain codes that can provide unequal error protection (UEP) property. To reduce the coding complexity and improve the performance of previous duplicate UEP schemes in the literature, we implement different rate low-density parity-check codes instead of the duplication process and replace the high-complexity classical robust soliton distribution by low constant average degree distributions with high intermediate symbol recovery rates. Asymptotic analysis of both the duplicate scheme and our scheme is obtained by deriving unequal density evolution formulas over the binary erasure channel. Compared to previous UEP schemes, our scheme has lower complexity. Experimental results show that our scheme can obtain better decoding performance, especially for smaller input information length. Comparison of peak signal-noise ratio performance for the scalable video coding shows that, for moderate input information length, our scheme can provide a better basic video quality at lower overhead, but needs larger overhead to achieve a high video quality.

Many real images are shot under hazy conditions and need to be processed for better quality. So far the majority of the dehazing methods try to faithfully invert the standard hazy image degradation model. In such an approach, not only is the basic problem theoretically unsolvable, but the parameters used in many methods are often hard to estimate accurately. In this paper we make two contributions. We first derive the standard haze model in the HSV color space, which is preferred over the traditional RGB color space for contrast enhancement due to its robustness to color distortion. Specifically, we show that under generally valid assumptions, the H channel is invariant to haze degradation. The S channel can be recovered when the ambient light and the medium transmission coefficients are estimated as in the traditional approach. Only the V channel takes the same form of the haze model as for the RGB channels. Secondly, we show that the standard image dehazing problem can be viewed as a type of image contrast enhancement. It is then possible to place a less stringent requirement on faithful parameter estimation in the framework of contrast enhancement. Based on such insight, we propose an algorithm that combines the image dehazing and contrast enhancement seamlessly. The major advantage of this new approach is that instead of degrading the processed image quality, the inaccurately estimated parameters with intentional bias tend to enhance the image contrast. Experimental results show that the new approach enjoys a large performance advantage over existing methods.

In this paper, we propose an efficient scheme to construct LT codes that can provide unequal error protection (UEP) property. We implement low rate LDPC codes instead of the duplication and utilize a low constant average degree distribution with high intermediate symbol recovery rates (ISRR) to improve the performance of more important bits (MIB). Asymptotic analysis of the proposed scheme is obtained by deriving unequal density evolution (DE) formulas over the binary erasure channel (BEC). Utilizing LDPC codes and the degree distribution with high ISRR in the proposed scheme, the UEP property is flexible and efficient. Compared with previous duplicate UEP-LT codes, mathematical analysis and simulation results demonstrate that the proposed UEP-LT codes have better performance.

Impulse noise is commonly encountered during image transmission and many methods have been proposed to remove it. Although it is now possible to recover the true image reasonably well, even under severe noise (90% pixel contamination), essentially all methods published so far follow the standard procedure of noisy pixel detection/classification and then noisy pixel value reconstruction, without any further processing. In this paper we show an interesting empirical discovery that the traditionally denoised image tends to have the estimation error with a Laplacian distribution, which makes it possible to add a postprocessing stage to denoise the traditionally obtained result with this new type of noise. We propose a practical algorithm within this new framework and experimental results show that superior results can be obtained over previously published methods.

Image noise level estimation is an important step in many image processing tasks such as denoising, compression and segmentation. Although recently proposed SVD and PCA approaches have produced the most accurate estimates so far, these linear subspace-based methods still suffer from signal contamination from the clean signal content, especially in the low noise situation. In addition, the common performance evaluation procedure currently in use treats test images as noise-free. This omits the noise already in those test images and invariably incurs a bias. In this paper we make two contributions. First, we propose a new noise level estimation method using nonlinear local surface approximation. In this method, we first approximate image noise-free content in each block using a high degree polynomial. Then the block residual variances, which follow chi squared distribution, are sorted and the upper quantile of a carefully chosen size is used for estimation. Secondly, we propose a new performance evaluation procedure that is free from the influence of the noise already present in the test images. Experimental results show that it has much improved performance than typical state-of-the-art methods in terms of both estimation accuracy and stability.

With the current rate of video data generation, there is an urgent need of automatic video content analysis for subsequent purposes such as summarization, retrieval and classification. And video shot boundary detection is usually the first step to segment a video clip into meaningful shots. Taking the processing speed into account, most state-of-the-art methods make use of the frame histogram to extract shot boundary characteristics. In this paper we propose a different approach with a novel metric, which essentially captures the observation that within any shot, a pixel value in any frame usually has a pixel value very close to it within a small neighborhood in an adjacent frame. It turns out that the proposed approach can make better use of frame structural content than the histogram approach. In addition, the proposed metric has a low computational complexity. We propose a video shot boundary detection algorithm based on the proposed metric for detecting both cut transition (CT) boundary and gradual transition (GT) boundary. Experimental results show that the proposed approach enjoys better detection rates over the state-of-the-art with competitive processing speed.

Although it has been recognized that different textual contents in an image need to be treated differently during accurate image interpolation, how to classify these contents well has been a difficult problem due to the inherent complexity in natural images. In this paper we propose an efficient image interpolation framework with a novel weighted surface approximation approach. The key is that the weighted mean squared error of the approximation can be converted to a continuously distributed probability of a pixel belonging to a local smooth region or a textural one, thus essentially making a soft pixel classification. In addition, the fitted local surface provides an estimate of the pixel value under the smooth region assumption. This estimate is then fused with the estimate from the texture region assumption using the previously obtained probability to yield the final estimate. Experimental results show that the proposed framework consistently improves over typical state-of-the-art methods in terms of interpolation accuracy while maintaining comparable computational complexity.

So far there are two separate approaches to removing salt-and-pepper noise: the median type filtering and the variational formulation. The first approach usually has fast speed, while the latter produces greatly improved result at much slower speed. In this paper we show that the variational approach can be approximated as a region growing process and propose a novel iterative algorithm that combines the strength of these two approaches. When viewed within a single iteration, the algorithm acts like a median type filter. When viewed across iterations, the filter achieves the region growing effect accomplished by the variational approach. Extensive simulations show that the proposed algorithm achieves the state of the art performance with the fastest speed published so far. The insight gained in this paper could have broader applications.

Code division multiple access (CDMA) has been proven to be a promising spread-spectrum multiple access technique due to its advantages such as anti-interference capability and implementation versatility. However, the problem of the receiver side synchronization at the sub-chip scale can significantly increase the decoding error rate, just as the carrier frequency offset (CFO) problem does to the OFDM systems. So far the use of faster sampling clock at the receiver analog-to-digital converter (ADC) is generally thought to improve the receiving performance and has been implemented in practical communication hardware such as the STEL-2000A CDMA chip. In this paper we show that the traditional use of such scheme can introduce interchip interference that causes an error floor irrespective of the sampling rate. We further propose a novel synchronization method that uses a digital filter bank to automatically seek out the best synchronization solution. Simulation results show that the proposed method performs almost as good as the case of perfect synchronization and far better than when there is a clock disparity. The proposed algorithm is also implemented in field programmable gate array (FPGA) and compared with the traditional approach to pave the way for its practical chip-level implementation.

Salt-and-pepper impulse noise is one commonly encountered noise type during image and video communication. So far the state of the art methods can reasonably restore images corrupted by salt-and-pepper noise whose level is up to 90%. We propose a novel quadratic type variational formulation of this noise removal problem. This approach first uses a simple yet fast method to eliminate all salt-and-pepper noise pixels as well as possibly some clean pixels, then the clean image is efficiently reconstructed from the remaining clean pixels by minimizing a carefully designed functional. Because the functional is quadratic type, fast unconditional convergence is guaranteed. Simulation results show that the proposed method outperforms previously published results and can tolerate noise level of as much as 95%.

The state of the art impulse noise removal methods make use of the noise variance, or equivalently the noise mixing probabilityp, and are iterative procedures (e.g., , ). However, so far there has been a lack of effective estimator forp. As a result, true values ofpare often used during simulation, which may not be practical. Furthermore, the optimal stopping criteria for the iterative algorithms have been elusive until recently. In a computationally heavy method is proposed for determining the optimal number of iterations. In this letter we make two contributions. We first develop a robust estimator forpby using the empirical observation that a natural image usually doesn't cover all pixel value range, then we design an efficient linear transformation to replace complicated computation of order statistics. Based on this estimatedpvalue, we further derive the formula for estimating the true image histogram, and use it to formulate a new efficient optimal stopping criterion during the iterative denoising process. This formulation has a simple interpretation of its optimality and yields improved denoising performance.

The subject of 2-D and higher dimensional object recognition finds widespread applications in areas such as image registration and pattern recognition. Radon transform is one technique used for efficient object matching (e.g., and ). However, so far as we know, no results have been obtained that solves the recognition problem completely in the projection domain due to coupling of transform parameters. We develop a novel method for such parameter decoupling and an improved phase correlation method for accurate practical shift estimation, resulting in a fast matching algorithm based on projection data only. Simulation results show that the proposed algorithm is much faster than similar state-of-the-art approaches such as that in with comparable estimation accuracy.

Histogram specification (equalization) is an important tool for tasks such as image enhancement and normalization. Although this problem has exact solution for continuous images, in the case of digital images, it is ill posed. Recently, an exact pixel ordering method based solely on local image intensity was proposed, yet still with some limitations, especially since it ignores the important image edge information. In this paper, we present a wavelet-based method that simultaneously achieves the exact histogram specification and good image enhancement performance. It does so through a carefully designed strict pixel ordering process, during which the wavelet coefficients are fine tuned for the image enhancement purpose. Compared to previous work, this approach takes into account not only local mean intensity values, but also local edge information. Other advantages include fast pixel ordering, good statistical models, and better image enhancement performance. Experimental results and comparison with state-of-the-art methods are presented.

Requantization is an important technique for variable-rate data compression and communication and optimal design for requantization is key to guaranteeing the minimization of the requantization noise. Recently, an iterative optimization algorithm for designing the requantization codebook was proposed. We develop further results regarding the acceleration of the algorithm and a better understanding of the problem of requantization. In particular, our result shows that the quantization and requantization processes can be decoupled into two nearly independent sequential stages, and that requantization does not need the original training vectors. All these features pave the way toward real-time variable-rate data compression and transmission over heterogeneous networks. Simulation results support the theory and the advantages of our method.