Statistical Multiplexing of Video Traffic and Data Traffic

This paper studies statistical multiplexing performance by input of video traffic and data traffic. The inputs have different Qos requirements such as loss and delay jitter. By applying a modified FBM model, we present methods to estimate effective bandwidth of the aggregated traffic. Simulations were performed to evaluate effective bandwidth. The comparison between the estimation and the simulation shows that the estimations can give correct data for the effective bandwidths in terms of our interests. The analysis of gain by using priority multiplexing also addresses proper Qos configuration for the inputs in order to achieve positive gains.

Mean shift based log-Gabor wavelet image coding

In this paper, we propose a sparse overcomplete image approximation method based on the ideas of overcomplete log-Gabor wavelet, mean shift and energy concentration. The proposed approximation method selects the necessary wavelet coefficients with a mean shift based algorithm, and concentrates energy on the selected coefficients. It can sparsely approximate the original image, and converges faster than the existing local competition based method. Then, we propose a new compression scheme based on the above approximation method. The scheme has compression performance similar to JPEG 2000. The images decoded with the proposed compression scheme appear more pleasant to the human eyes than those with JPEG 2000.

Resynchronization and remultiplexing for transcoding to H.264/AVC

H.264/MPEG-4 AVC standard appears highly competitive due to its high efficiency, flexibility and error resilience. In order to maintain universal multimedia access, statistical multiplexing, or adaptive video content delivery, etc., it induces an immense demand for converting a large volume of existing multimedia content from other formats into the H.264/AVC format and vice versa. In this work, we study the remultiplexing and resynchronization issue within system coding after transcoding, aiming to sustain the management and time information destroyed in transcoding and enable synchronized decoding of decoder buffers over a wide range of retrieval or receipt conditions. Given the common intention of multiplexing and synchronization mechanism in system coding of different standards, this paper takes the most widely used MPEG-2 transport stream (TS) as an example, and presents a software system and the key technologies to solve the time stamp mapping and relevant buffer management. The solution reuses previous information contained in the input streams to remultiplex and resynchronize the output information with the regulatory coding and composition structure. Experimental results showed that our solutions efficiently preserve the performance in multimedia presentation.

Multiple MIPS 4Kc cores based interrupt controller design and its implementation on HDTV SoC platform

A multiple MIPS 4Kc processor cores based interrupt processing system is introduced. The interrupt controller plays a key role in the high definition television （HDTV） system-on-a-chip （SoC） platform, especially when it is a multiple processor system. Based on a general introduction to the whole HDTV SoC platform, a layered interrupt controller and its implementation are discussed in detail. The proposed scheme was implemented in our FPGA verification board. The results indicate that our scheme is reliable and efficient. Meanwhile, as a functional intellectual property （IP）, the interrupt controller has reusability and expandability with the layered structure.

Compressed Multi-image Reconstruction Based on Quantization Noise Distribution

Because of the quantization noise introduced during the compression,super-resolution reconstruction(SRR)techniques are complicated for the compressed images.This paper aims to incorporate the prior knowledge of discrete cosine transform(DCT)coefficients into modeling the quantization noise.The spatial covariance matrix of the quantization noise is estimated by utilizing the Laplacian distribution of the alternating current(AC)coefficients.After estimating the spatial joint covariance of overall noises for the imaging system,we propose a general Bayesian framework to enhance the resolution for compressed images.Experiments demonstrate the effectiveness of the proposed algorithm and show the superiority to previous methods in objective and subjective aspects.

An efficient training sequences strategy for channel estimation in OFDM systems with transmit diversity

This paper deals with channel estimation for orthogonal frequency-division multiplexing (OFDM) systems with transmit diversity. Space time coded OFDM systems, which can provide transmit diversity, require perfect channel estimation to improve communication quality. In actual OFDM systems, training sequences are usually used for channel estimation. The authors propose a training based channel estimation strategy suitable for space time coded OFDM systems. This novel strategy provides enhanced performance, high spectrum efficiency and relatively low computation complexity.

A compression tolerant scheme for image authentication

Image authentication techniques used to protect the recipients against malicious forgery. In this paper, we propose a new image authentication technique based on digital signature. The authentication is verified by comparing the features of the each block in tested image with the corresponding features of the block recorded in the digital signature. The proposed authentication scheme is capable of distinguishing visible but non-malicious changes due to common processing operations from malicious changes. At last our experimental results show that the proposed scheme is not only efficient to protect integrity of image, but also with low computation, which are feasible for practical applications.

Quantization of wavelet packet audio coding

Abstract The method of quantization noise control of audio coding in the wavelet domain is proposed. Using the inverse Discrete Fourier Transform （DFT）, it converts the masking threshold coming from MPEG psycho-acoustic model in the frequency domain to the signal in the time domain; the Discrete Wavelet Packet Transform （DWPF） is performed; the energy in each subband is regarded as the maximum allowed quantization noise energy. The experimental result shows that the proposed method can attain the nearly transparent audio quality below 64kbps for the most testing audio signals.

Robust Video Stabilization Based on Motion Vectors

This paper proposes a new robust video stabilization algorithm to remove unwanted vibrations in video sequences. A complete theoretical analysis is first established for video stabilization, providing a basis for new stabilization algorithm. Secondly, a new robust global motion estimation (GME) algorithm is proposed. Different from classic methods, the GME algorithm is based on spatial-temporal filtered motion vectors computed by block-matching methods. In addition, effective schemes are employed in correction phase to prevent boundary artifacts and error accumulation. Experiments show that the proposed algorithm has satisfactory stabilization effects while maintaining good tradeoff between speed and precision.

A fast block-matching algorithm based on variable shape search

Block-matching motion estimation plays an important role in video coding. The simple and efficient fast block-matching algorithm using Variable Shape Search (VSS) proposed in this paper is based on diamond search and hexagon search. The initial big diamond search is designed to fit the directional centre-biased characteristics of the real-world video se- quence, and the directional hexagon search is designed to identify a small region where the best motion vector is expected to locate. Finally, the small diamond search is used to select the best motion vector in the located small region. Experimental results showed that the proposed VSS algorithm can significantly reduce the computational complexity, and provide competitive computational speedup with similar distortion performance as compared with the popular Diamond-based Search (DS) algorithm in the MPEG-4 Simple Profile.

Spatio-temporal Compensation Based Object Detection for Video Surveillance Systems

Moving object detection in video surveillance is an important step. This paper addresses an automatic object detection algorithm based on spatio-temporal compensation for video surveillance. Temporal difference of the pairs of two frames with a k-frame distance is utilized to obtain coarse object masks. Usually, object regions in these coarse masks have discontinuous boundaries and some holes. Region growing with the distance constraint is proposed to compensate these coarse object regions in spatial domain, followed by filling holes. The added distance constraint can prevent object regions from growing infinitely. The proposed filling holes method is simple and effective. To solve the temporarily stopping problem of moving objects, temporal compensation is proposed to compensate the object mask by utilizing temporal coherence of moving objects in temporal domain. The proposed detection algorithm can extract moving objects as completely as possible. Experimental results have successfully demonstrated the validity of the proposed algorithm.

Identification of various food residuals on denim based on hyperspectral imaging system and combination optimal strategy

As the science and technology develop,crime methods and scenes have become increasingly complex and diverse.Trace evidence analysis has become amore and more important criminal investigation technology and liquid is the main form of trace evidence.Food can provide not only energy,but clues to solve crimes.In this study,we build a hyperspectral imaging system to detect liquid residue traces,including apple juice,coffee,cola,milk and tea,on denims with light,middle and dark colors.The obtained hyperspectral images are first subjected to spectral calibration and hyperspectral data pretreatment.Subsequently,Partial Least Squares(PLS)is applied to select the informative wavelengths from the preprocessed spectra.For modeling phase,the combination optimal strategy,support vector machine(SVM)combined with random forest(RF),is developed to establish classification models.The experimental results demonstrate that the combination optimal model can achieve TPR,FPR,Precision,Recall,F1,and AUC of 83.5%,2.30%,79.7%,83.5%,81.6%,and 94.7%for classifying fabrics contaminated by various food residuals.With respect to the classification of liquid and fabric types,the combination optimalmodel also yields satisfactory classification performance.In future work,wewill expand the types of liquid,and make appropriate adjustment to algorithms for improving the robustness of classification models.This research may play a positive role in the construction of a harmonious society.

A Regularized Super Resolution Algorithm for Generalized Gaussian Noise

In this paper,an iterative regularized super resolution (SR) algorithm considering non-Gaussian noise is proposed.Based on the assumption of a generalized Gaussian distribution for the contaminating noise,an lp norm is adopted to measure the data fidelity term in the cost function.In the meantime,a regularization functional defined in terms of the desired high resolution (HR) image is employed,which allows for the simultaneous determination of its value and the partly reconstructed image at each iteration step.The convergence is thoroughly studied.Simulation results show the effectiveness of the proposed algorithm as well as its superiority to conventional SR methods.

Structured Computational Modeling of Human Visual System for No-reference Image Quality Assessment

Objective image quality assessment(IQA)plays an important role in various visual communication systems,which can automatically and efficiently predict the perceived quality of images.The human eye is the ultimate evaluator for visual experience,thus the modeling of human visual system(HVS)is a core issue for objective IQA and visual experience optimization.The traditional model based on black box fitting has low interpretability and it is difficult to guide the experience optimization effectively,while the model based on physiological simulation is hard to integrate into practical visual communication services due to its high computational complexity.For bridging the gap between signal distortion and visual experience,in this paper,we propose a novel perceptual no-reference(NR)IQA algorithm based on structural computational modeling of HVS.According to the mechanism of the human brain,we divide the visual signal processing into a low-level visual layer,a middle-level visual layer and a high-level visual layer,which conduct pixel information processing,primitive information processing and global image information processing,respectively.The natural scene statistics(NSS)based features,deep features and free-energy based features are extracted from these three layers.The support vector regression(SVR)is employed to aggregate features to the final quality prediction.Extensive experimental comparisons on three widely used benchmark IQA databases(LIVE,CSIQ and TID2013)demonstrate that our proposed metric is highly competitive with or outperforms the state-of-the-art NR IQA measures.

Unequal decoding power allocation for efficient video transmission

We present an unequal decoding power allocation （UDPA） approach for minimization of the receiver power consumption subject to a given quality of service （QoS）, by exploiting data partitioning and turbo decoding. We assign unequal decoding power of forward error correction （FEC） to data partitions with different priority by jointly considering the source coding, channel coding and receiver power consumption. The proposed scheme is applied to H.264 video over additive white Gaussion noise （AWGN） channel, and achieves excellent tradeoff between video delivery quality and power consumption, and yields significant power saving compared with the conventional equal decoding power allocation （EDPA） approach in wireless video transmission.

A Bayesian Super Resolution Algorithm Based on Synthetic Gradient Distribution

A novel Bayesian super resolution (SR) algorithm based on the distribution of synthetic gradient is proposed. The synthetic gradient combines prior information in horizontal, vertical, and diagonal directions. Its distribution is modeled as a Lorentzian function and regarded as a new image model which can sufficiently regularize the ill-posed algorithm and preserve the edges in the reconstructed images. The graduated nonconvexity (GNC) optimization is employed to guarantee the convergence of the proposed Lorentzian SR (LSR) algorithm to the global minimum. The performance of LSR is compared with conventional algorithms, and experimental results demonstrate that the proposed algorithm obtains both subjective and objective gains.

Mode decision for H.264 down-scaling video transcoding

Although the coding modes of H.264 coded video would be changed by the transcoding process of spatial resolution reduction, there exists good correlation in prediction modes and prediction directions between input and output video. In this paper, we first introduce a new spatial resolution reduction transcoding architecture of intra coded frames where the distortion can be calculated directly in compression domain. We then propose a fast mode decision algorithm in which only a small part of rate distortion optimization （RDO） calculation is needed for mode decision. For 4×4 luma block, the proposed scheme has average 21.3% computation saving, compared to the cascaded pixel-domain transcoding scheme with the fast intra mode decision algorithm proposed in JVT-G013. For 16×16 luma block, RDO calculation is completely avoided in our scheme while the scheme in JVT-G013 needs 2 RDO calculations. Experimental results show that our scheme outperforms that of JVT-G013 in terms of significantly computasavings with negligible loss of PSNR

Macroblock-Level Rate Control Scheme for Video Encoders

Rate control plays an important role in video coding. An algorithm is proposed by Tsai et al in which macroblock’s coding order is rearranged to improve the coding quality. However, the complexity is very high due to the change of macroblock’s coding order. In this paper, a macroblocl-level rate control algorithm is proposed which recalculates the quantization parameter of each macroblock based on its significance. Simulation results show that the proposed algorithm not only achieves 0.1-0.6 dB in peak signal-to-noise ratio (PSNR) but also reduces 33%-55% total encoding time compared to Tsai’s algorithm.

Auxiliary Diagnosis of Papillary Thyroid Carcinoma Based on Spectral Phenotype

Thyroid cancer,a common endocrine malignancy,is one of the leading death causes among endocrine tumors.The diagnosis of pathological section analysis suffers from diagnostic delay and cumbersome operating procedures.Therefore,we intend to construct the models based on spectral data that can be potentially used for rapid intraoperative papillary thyroid carcinoma(PTC)diagnosis and characterize PTC characteristics.To alleviate any concerns pathologists may have about using the model,we conducted an analysis of the used bands that can be interpreted pathologically.A spectra acquisition system was first built to acquire spectra of pathological section images from 91 patients.The obtained spectral dataset contains 217 spectra of normal thyroid tissue and 217 spectra of PTC tissue.Clinical data of the corresponding patients were collected for subsequent model interpretability analysis.The experiment has been approved by the Ethics Review Committee of the Wuhu Hospital of East China Normal University.The spectral preprocessing method was used to process the spectra,and the preprocessed signal respectively optimized by the first and secondary informative wavelengths selection was used to develop the PTC detection models.The PTC detection model using mean centering(MC)and multiple scattering correction(MSC)has optimal performance,and the reasons for the good performance were analyzed in combination with the spectral acquisition process and composition of the test slide.For model interpretable analysis,the near-ultraviolet band selected for modeling corresponds to the location of amino acid absorption peak,and this is consistent with the clinical phenomenon of significantly lower amino acid concentrations in PTC patients.Moreover,the absorption peak of hemoglobin selected for modeling is consistent with the low hemoglobin index in PTC patients.In addition,the correlation analysis was performed between the selected wavelengths and the clinical data,and the results show:the reflection intensity of selected wavelengths in normal cells has a moderate correlation with cell arrangement structure,nucleus size and free thyroxine(FT4),and has a strong correlation with triiodothyronine(T3);the reflection intensity of selected bands in PTC cells has a moderate correlation with free triiodothyronine(FT3).

Progressive refinement for robust image registration

A new image registration algorithm with robust cost function and progressive refinement estimation is developed on the basis of direct method (DM). The robustness lies in M-estimation to avert larger local noise and outliers. Moreover, the progressive refinement model estimation under the multi-resolution framework, where the initial parameter values of coarse level are estimated by Mellin transformation, is adopted so as to get global optimization and reduce search space. Experiments show that the proposed algorithm greatly extends the scope of the conventional DM of image registration.