Audio blind watermarking scheme combined DWT DCT QR

In order to protect the copyright of digital audio and video in Internet, we propose a novel audio blind watermarking scheme combined discrete wavelets transform, discrete cosine transform, QR decomposition and audio characteristics. In this algorithm, the audio are split into blocks, and each block are decomposed on two dimensional discrete wavelet transform (DWT), then the approximate sub band coefficients are decomposed on discrete cosine transform (DCT), and the first quarter of the DCT coefficients are decomposed on QR decomposition and get a triangle matrix. At last, the watermarking information is embedded into the triangle matrix. The experiments show that the algorithm can get better balance between transparency and robustness of watermark, and it has strong robustness against the common audio signal processing such as additive white Gaussian noise, re sampling, re quantization, low pass filter, MP3 compression and cutting replacement.     

林业专题图数字水印版权保护研究

针对林业专题图制作过程复杂、成本高、迫切需要进行版权保护的问题,分析了JPEG图像压缩的不变属性,在此基础上提出了一种DCT域数字水印算法,该算法将水印信息嵌入到JPEG格式的林业专题图像的中高频系数当中。结果表明：因为水印信息是嵌入在中高频系数上,所以不仅嵌入的水印信息不可感知性良好,而且水印的嵌入容量也有所增加;即使对加密林业专题图进行有损压缩时使用高于预设值的质量因子,此算法仍然能够完全一致地提取出嵌入的水印信息,并且对各种噪声干扰都具有较好的鲁棒性。     

Binary phase only filter based watermarking combined with discrete wavelet transform

The binary phase only filter (BPOF) based digital image watermarking combined with a discrete wavelet transform (DWT) was proposed. Firstly, the DWT transform was applied to an image. Then, the discrete Fourier transform (DFT) was applied to the low frequency subband of the DWT transform. Next, the phase information of the DFT was binarized to obtain the BPOF, which would be taken as the watermark and embedded into the corresponding magnitude. Compared to applying the DFT transform to the whole image and embedding the BPOF watermark in the entire frequency range, or in the low frequency range, the watermark robustness in JPEG compression is improved significantly while maintaining the watermark imperceptibility and detection efficiency. This method may be used to demonstrate the authenticity and integrity of an image. The simulation experiments demonstrate the efficiency of this method.     

Image Watermarking Technique Based on the DCT Transform and the Edge Detection

As a main application of information hiding technique to multimedia copyright protection, the digital watermarking technique, which embeds an invisible signal (watermark) to the original multimedia data (image, audio, and video), has been proposed as a viable solution for the copyright protection and authentication of multimedia data in a networked environment since it makes it possible to identify the author, owner, distributor or authorized consumer of multimedia data. So it has drawn extensive attention in recent years and has been widely applied. A new image watermarking scheme is proposed by incorporating the DCT transform and an edge detection technique. Experimental results show that this method is more robust against the common signal processing attacks than the traditional Cox's method.     

Watermarking algorithm based on statistical property

A novel robust blind watermarking algorithm in DCT domain was proposed. The DCT coefficients are arranged in a specific way, and then the watermark is embedded based on odd even difference for optimal embedding. This algorithm works in a new pattern other than block division and search of the DCT coefficients in the middle frequencies in the traditional pattern, which can overcome the problem in searching suitable coefficients in the middle frequencies for watermark embedding and the small embedding capacity. Experimental results show that the algorithm is simple and good at perceptual transparency as well as robustness against noise and JPEG compression.     

A Medical Image Quantization Method Suited for PACS

Image compression is very important in picture archiving and communication system(PACS). The author studied the statistical distribution of image wavelet subimage coefficients and concluded that the distribution of wavelet subimage coefficients is similar to that of Laplasian distribution. On the other hand, in image reconstruction, the coefficient with different amplitude owns different weight, and different accuracy can be applied to different coefficients according to their different weight. Then, the author has designed a image quantization encoding scheme for PACS. In this scheme, they selected the sample-standard-deviation of coefficients in every subimage as the quantization threshold, and accurately encoded those coefficients with higher weight. Also, this algorithm utilized the visual character of human. The test has proved that the main advantages of this method are the simplicity in computing and predictable encoded coefficients, and a high compression efficiency can obtain too.     

Single sample face recognition based on B-spline and image gradient

A single sample face recognition algorithm based on B-spline and image gradient is proposed. Image gradient method for face recognition has advantage of illumination invariant. But the recognition rate will be greatly decreased when the image contains noise which will seriously influence gradient information. Traditional methods to reduce noise smooth image at the same time and image gradient reorganization rate will be reduced. As the B-spline filter has the feature which can adjust the order, B-spline filter with different orders can be selected according to the image noise value to minimize noise while preserve image gradient information. Experiments prove that using B-spline and image gradient algorithm can achieve a better recognition rate than traditional filtering method on single sample face recognition problem.     

An optimal equivalent subchannel algorithm for low voltage power line communication system

According to the low voltage power line channel features,combined with adaptive bit loading and power allocation algorithm,an optimal equivalent subchannel algorithm is proposed.Under the condition of fixed amount of transmitted bits and the minimum transmitted power model,impact of optimal equivalent subchannel on the system performance is studied.The optimization algorithm sorts subchannels with similar transmission features into groups,thus the signaling load can be reduced,and the spectrum can be better used.The simulation results show that system can get lower bit error rate(BER) by adopting optimal equivalent subchannel algorithm.     

Illumination compensation of color video for driver fatigue monitor

A robust method to compensation illumination of driver video color image to enhance image quality has been proposed. Firstly, the original image is decomposed into high-frequency image and low-frequency image based on spatial image decomposition. The illumination information is completely contained in the low-frequency image. Next, low-frequency image is compensated by using Retinex approach which can brighten the darker area. Then, the compensated low-frequency image and high-frequency image are synthesized. Finally, new image is obtained by illumination balance of synthesized image. Results show that the algorithm has better results compared with the traditional algorithm.     

A borrowed address based on efficient distributed address assignment algorithm in ZigBee networks

The borrowed address algorithm can solve the orphan problem made by the restriction of threshold of children when DAAM(distributed address assignment mechanism) assigns the addresses for the nodes in ZigBee networks. The exited borrowed algorithms only can increase the success ratio of address assignment, but their overhead and times spend on founding net cost much. To address the problem, we propose an efficient borrowed address assignment algorithm EDAA BA (efficient distributed address assignment algorithm based on borrowed address). To control the overhead and time spend on founding net, it borrows addresses firstly from the descent nodes in same branch for the orphan router nodes, adds the mechanism of immediate reply of borrowed message, and assigns the remainder addresses unused by DAAM in the 16 bits addresses for the orphan end devices. Theoretical and simulation analyses show that our algorithm outperforms DAAM and two of its present improvement algorithms in terms of the overhead and time spent on forming network, under the premise of better success rate of address assignment.     

Automatic head recognition by integrating mean shift with multiple features

To improve the head detection accuracy in video sequences captured with fixed vertical monocular camera, a novel method of head recognition based on mean shift and multiple features is proposed. Firstly, mean shift based image segmentation algorithm with color information and spatial information is suggested to derive the candidate head components in images. Furthermore, by referring to two features that the contour of human head regions are approximate round and the hair color distribution is clustered, the evaluation models based on the contour information and inside color information of candidate head components are presented for head recognition. The experimental results show that the proposed algorithm can effectively reduce the light interfere and eliminate fake target whose color information is similar to hair color distribution. The detection rate for static images can reach about 89.4%.     

Low complexity rate control scheme for real-time video coding

Considering the deficiency of H.264 rate control scheme and the coding efficiency requirements for real-time encoding,a low complexity rate control algorithm in MB-layer is proposed..MAD is predicted by combining motion vector and the temporal and spatial correlation to reduce the complexity of G012-MAD prediction algorithm from linear regression to the complexity of the weighted average. For the rate-distortion model,the linear model with low computational complexity instead of the quadratic fitting algorithm in G012 is used,while the possibility of is discussed combining with SSIM. For the fluctuation of the PSNR of the coded image with large movement,the determination of QP in GOP layer and the bit allocation of P frame are improved,and the current quantization parameter is adjusted by making full use of the information of coded MBto control the rateeffectively. The experimental results show that the algorithm is more effective in resisting the fluctuation of PSNR for smoothly output sequence,and can decrease computational complexity.     

储粮害虫图像的灰关联分析方法研究

对储粮害虫进行实时监测是非常必要的，其中基于图像识别的检测是目前研究较多的一种方法。本文提出了一种基于灰关联分析的图像边缘检测的新算法，并将其应用于储粮害虫图像的边缘检测，计算表明，该方法能有效地检测粮虫图像的边缘信息。     

Segmentation algorithm for urinary sediment image combiningwavelet transform and 2D Maximum entropy threshold

In order to solve the problem that urine sediment visible components cannot be segmented effectively because of complex components, complicated defocusing in image and poor discrimination between object and background, a method based on combination algorithm wis designed to segment urine sediment. The wavelet transform wis used to erase the effect of defocusing. Then morphology wis utilized to get the subimages that include the particles. The segmentation method combining the wavelet transform based segmentation and the two dimensional entropy threshold based segmentation wis employed to segment urine sediment visible components. Experimental results show that the proposed method can segment urinary sediment images effectively and precisely.     

Rate control using the rate-distortion properties of the foreground and background regions

In order to improve the rate control performance of video encoder, video frames are segmented into two regions according to the video content, i.e. the foreground region containing the main moving objects and the background region containing the relatively static scene. The rate-distortion properties of the two regions are analyzed. Based on the analyses, a novel region-based macroblock layer rate control method for H.264/AVC is proposed. For the bit allocation process, more target bits can be assigned to the important macroblocks by employing an adaptive factor determined by the region information. Then, the quantization parameter (QP) is computed by using the quadratic model with different parameters related to the two regions. The computed QP is further adjusted to suppress the coding quality fluctuation between the two regions, as well as the quality fluctuation among the consecutive macroblocks. Finally, the rate control parameters are updated after coding the current macroblock. Experimental results demonstrate the mismatches between the actual coding bit-rates and the target bit-rates are reduced, while the coding performance of the encoder and the subjective visual quality of the reconstructed video are improved.     

Multi scale and multi orientation features for image de noising

An adaptive algorithm for image de noising is proposed based on the multi scale and multi orientation features. The coefficients in different scales and different directions are obtained by image decomposition using the nonsubsampled contourlet transform. Then thresholds functions are adaptively set with these coefficients. The texture of the image information is introduced by using the mean of decomposition scale and the energy of regional. The greater the energy, the more information of the texture while the same decomposition scales, the smaller the threshold is set. On the contrary, the greater the threshold is set. After the de noising and then reconstruction of these coefficients, image de noising is implemented. Compare to the wavelet transform threshold and contourlet transform threshold, the nonsubsampled contourlet transform pick up the image detail better and improve the quality of the image.     

A fast image segmentation algorithm with local C V active contour model

For the typical C V Algorithm, there exists the weakness of requiring multi iterative operations and long time computation to deal with large size image. Based on the analysis upon the relationship between the image size and the initialized approaching image with the number of iterations and computing time to obtain the stead results, an improved local C V image divisional algorithm based on the segmentation of threshold value and the connected domain labeled algorithm to deal with large size image is proposed. The OTSU method is used to divide the threshold value of image to reach the goal of label and local segmentation of image through the fast non recursion algorithm of connected domains. The segmented pieces and the result of its segmentation are used as the initialized approaching image of the C V algorithm model. Compared with the classical C V algorithm, the analysis and simulated result indicates that the improved C V algorithm reaches the steady solution quickly with fewer times of iterations. The proposed method can handle large size image with profound contour details quickly and effectively.     

Lossless data hiding with ranks transform

For taking full advantage of carrier image's pixel correlation, a lossless data hiding algorithm with ranks transform (LDHA_RT) is proposed. Along with three directions of column, row and diagonal, the pixel difference images and their histograms are calculated one after another. The obtained pixel difference histograms are shifted to construct the redundancy spaces, so that the hiding information could be embedded in a designated position. After the hiding data is extracted, the carrier image could be recovered reversibly without any additional information. The experiments demonstrate that the PSNR/bits of the LDHA_RT is the lowest or close to the lowest, compard with these algorithms mentioned in the references. This implies that the LDHA_RT has large capacity of hiding data under a good image quality.     

A floating threshold quantization algorithm to suppressmixing interferences in partial discharge

Due to the obvious difference of energy distribution frequencies from partial discharge (PD) signal and its mixing interferences (white noise and narrow brand), we uses the characteristic that node decomposition coefficients of wavelet packet transform can effectively show the energy change of signals to build up a floating threshold quantization algorithm (FTQA) varying with the noise energy of PD decomposition coefficients. It makes the node thresholds under the optimal base various with the noise strength in decomposition coefficients to self adaptively reality the choice of optimal threshold to finely partition PD decomposition coefficients. For simulated and real PD signals with mixing interferences, the conditional global threshold quantization algorithm (GTQA) and the proposed floating threshold quantization algorithm are employed to suppress the mixing interferences in PD signals and compared, and the results show that the proposed algorithm has the stronger suppression ability to mixing interference on PD signal and keeps perfect PD waveform via suppression.     

Disign and Implementation of Speaker Recognition System Based on FPGA

Aiming at the shortcoming of low recognition and training speed in embedded speaker - recognition system based on DS Phard-core processor, a new scheme of system based on FPGA and vector quantization principle is presented. In the speaker-recognition system based on the vector quantization and generation algorithm, a fitness parallel process hardware structure is presented by the scheme which consumes much less time than software processing while getting the fitness. The test shows that the system uses this method obtains both high recognition rate and higher speed in training and recognition.