Earticle

SAR image matching as a key technology is valuable in theory and practicality. Recent years, wide attention has been paid to cross-correlation algorithm and Hu invariant moments algorithm in the image matching field. The paper introduces concept of the two methods, and takes samples to analyze them to summarize the merits and faults. These two methods are not so effectual for SAR image matching. So based on the traditional Hu invariants moments algorithm, a new kind of improved Hu invariant moments algorithm is proposed. The experimental results prove that the improved method is stable and reliable under some variations for rotation in SAR image matching.

Facial expressions analysis is a vital part of the research in human-machine interaction. This chapter introduces an automatic recognition system for facial expression from a front view human face image. Obtaining an effective facial representation from initial face images is an essential phase for strong and efficient facial expression system. In this chapter we have developed a facial expression analysis system. Firstly we have tried to evaluate facial analysis base on Active shape model (ASM). In order to detect the face images, Adaboost classifier and Haar-Like feature has been adopted to achieve face detection and tracking. The ASM then has been automatically initiated in the detected face image. Then, discriminates and reliable facial feature points has been extracted applying ASM fitting technique. The geometric displacement among the projected ASM feature points coordinates and the mean shape of ASM were used to evaluate facial expression. Using support vector machine (SVM) classifier, the obtained results has reached a recognition rate of 93 %.

A novel mechanism model is proposed for the difference between the lake water areas derived from the remote sensing images at high spatial resolution and low spatial resolution. The areas estimated by images at different spatial resolution are calculated theoretically for the elliptical (including circular), triangular and hexagonal areal ground features. Based on the ideal cases, the quantitative rule between the two kinds of area is summarized. According to the theoretical results, the mechanism model is established for the error of the lake surface area. Especially, the surface area of Ebinur Lake derived with SPOT/VEGETATION (VGT), Landsat EM/ETM+ and MODIS are employed to validate the effectiveness and applicability of this model. With the help of the linear regression method, this model can be used to estimate the error between the Ebinur Lake surface areas derived from satellite images at different spatial resolutions, and promote the lake surface area from low spatial resolution imagery. Consequently, this model is helpful for estimating the lake surface area more accurately in a large region, and its effectiveness can be improved with more historical data.

Healthcare institution that handles a number of patients, opinions is often sought from different experts. It demands the exchange of the medical history of the patient among the experts which includes the medical images, prescriptions and electronic patient records (EPR) etc. In order to reduce storage and transmission cost, data hiding techniques are used to embed patient information with medical images. In medical imaging applications, there are stringent constraints on image fidelity that strictly prohibit any permanent image distortion by the watermarking or data hiding. Authenticity is another important aspect in medical image watermarking. This paper proposed modified difference expansion watermarking using LSB replacement in the difference of virtual border for data hiding in medical images. The Class Dependent Coding Scheme (CDCS) is used to encode the EPR data so that embedding capacity can be increased. The image hash is calculated using MD5 to provide authentication. Experimental results show that proposed scheme provide us large data hiding capacities along with very high PSNR values as compared to earlier EPR data hiding techniques.

Based on the analysis of the shortcomings of the current KLT (Kanade-Lucas-Tomasi) algorithm, a new, self-adaptive tracking algorithm is devised. Through introducing a kind of filtering mechanism, interference from surrounding noise and light on the tracked target is reduced. In order to reduce the error in tracking, a method based on forward-backward error approximation is utilized. Since such approximation reduces the number of visual feature points on the target, when the target changes in its shape tracking failure may easily result. In order to prevent this possible failure, a mechanism that appends the feature points is introduced. Experimental comparisons show that the said algorithm is clearly more effective than other similar algorithms.

In order to solve the problem aliasing of Contourlet transform, an Image Fusion Watermarking Algorithm based on NSCT and constrained NMF is proposed in this paper. Firstly, the image is decomposed into several layers. Secondly, the Two-dimension matrix is obtained by low frequency coefficients and watermarking signal. Finally, the watermarking is embedding by constrained NMF. The blind detection is used, and the algorithm can achieve a good imperceptibility. Experimental results show that the algorithm has good robustness against attacks of common image processing.

Now a day’s information security is most wanted achievement, and its need in the communication. The intends of this entitle work is wished-for an image encryption and decryption. Image encryption is an effective approach for providing security and privacy protection for medical images. This feature is especially very useful for multi medical images where the images can be making as one. In this work applies rational operators in between two images at multi levels to make multi key image to improve security levels. This work mainly approaches two operations that are OR and XOR logical operations.

To deal with the noise from rotating machinery vibration signal and analysis the fault signal, a rotating machinery fault diagnosis method based on enforced de-noising and modified EMD is proposed. Firstly, fault signals were de-noised by the wavelet with enforced threshold in order to filter the noises in the high frequency, and then the EMD method used to decompose the fault signals into a finite number of stationary intrinsic mode function (IMFs), then the linear correlation coefficient between two sets of data is proposed to select the useful IMF. In order to restrain the endpoint effect of EMD, in this paper, the cosine window function is employ to the fault signals, and then the envelope error of the fault signals is controlled at the both endpoints of the vibration signals. Experimental results shows: this proposed method can extract the fault information effectively, with overcoming the drawbacks of EMD well.

The advantages of micro strip patch antennas have made them a perfect candidate for use in the local area network (WLAN) applications. This paper represents the new design of micro strip patch antenna for China Multimedia Mobile Broadcast (CMMB). In the designed micro strip patch antenna, we applied different techniques to achieve the resonant frequency of 2.7GHz with the bandwidth of 300MHz. Unique slots and conventional slits are introduced in such a way that not only enhanced the bandwidth but also provided us with the fruitful result of -74.78dB return loss. Its minor non-symmetry led to dual frequency i.e. 5.60GHz with the bandwidth of 600MHz and return loss of -15.41dB. Further the details regarding to the dimensions of the patch, ground and simulations are elaborated.

Random demodulation based compressive sampling technique is a novel approach that it can break through the Shannon sampling theorem for the sparse signal capturing. A major challenge in the random demodulation based sampling system is the random sequence generation. In this paper, we introduce an approach to generate the high-speed random sequence that meets the incoherence of compressive sampling. The proposed technique employs a field programmable gate array (FPGA). First, the random sequence is parallel stored in the memory of FPGA, and it is read out byte by byte using a low speed clock. Second, the low-speed byte sequence is converted to a high-speed bit sequence by a circuitry. This proposed approach can program the random sequence dynamically without making any change to the circuitry system. Experimental results indicate that, the random sequence generated by the proposed approach is feasible to sensing the signal, and the constructed system can compressively sample and reconstruct the sparse signal.

Level set methods have been widely used in image processing specially in image segmentation. This paper presents a new region based active contour model in a variational level set formulation for segmentation of real world images in the presence of intensity in-homogeneity and noise. In this paper, we derive a local intensity clustering property in the image domain with better distance regularization function. The level set methods sometimes develop irregularities during its evolution state, which may cause numerical complexity and destroy the stability of evolution. This distance regularization function is able to maintain the desired shape of level set function smoothly and eliminates the need of re-initialization of LSF. The local clustering criterion function is defined for image intensities in neighborhood of each point. Now, this local clustering criterion of point is then integrated with respect to the neighborhood of entire points for global clustering criterion of image segmentation. In which bias function is also evaluated to intensity inhomogeneity correction. Implementation of our method shows that, it is more robust to initialization, and more accurate than conventional model.

Hausdorff distance is a common image registration method that is based on the edge features in the image. Theoretically, using Hausdorff distance method, the rotation, scaling and translation factors of the image can be obtained by searching the four-dimensional space that includes one rotation factor, one scaling factor and two translation factors. However, each additional factor means that the dimension degree of searching space is increased one more. The searching speed will be greatly reduced with the increasing of dimension degree. This paper presents a new image registration algorithm that combines the Hausdorff distance with the angle transform. Using this method, the translation factors can be obtained by Hausdorff distance and the angle transform can be directly computed. There is three step for finishing the image registration through the new method. Firstly, discrete Canny edge detection is used on the image. Secondly, the linear features of the image is directly used to calculate the rotation angle parameter. Then Hausdorff distance is used searching in two-dimensional space for getting the translation parameters of the image. Simulation results verified the correctness and validity of the paper method.

Through the application of theories and principles of human-computer interaction design system, deficiencies on interaction of digital music player are analyzed; and through the analysis on cognitive features for interaction, hierarchy of needs, and emotional information of target users, the key factors and the relevant data that influence the objective of interaction design are researched, thus facilitating the modern digital music products design to follow up users’ physical and mental requirements, balancing the interaction design between human, computer and environment, bringing users more convenient and effective operation and experience, and providing more important theoretical thought and implementation process for humanized design.

One of the powerful optimization tools that has been exploited in the computer world are nature inspired algorithms (NIAs), they are also used to solve problems in the computer programming world. For many years new algorithms have been developed regarding computer science and engineering communities such algorithms concentrates on NIAs which has proven their capabilities in many aspect, in some situation rapid solutions are needed to solve some problems these algorithms provides a versatile robust solutions for many of these situations. This paper presents a watermark inclusion based on a recently presented nature inspired algorithm to enhance the digital image watermarking procedure to be used for copyright protection. The nature inspired algorithm in focus is used to perfectly identify optimal positions in the discrete wavelet transform domain (DWT) for watermark inclusion in the gray scale image, The obtained results are shown in the experimental results section clarifying the superiority of using the algorithm in focus for the watermarking technique, In addition, showing how the algorithm optimum positions are obtained with lowest effect to the PSNR value of the produced watermark included images.

In this paper we have combined both cryptography and steganography techniques. This provides the higher level of secure system in which the secret information can be transferred over any unsecured communication channel and to overcome the threat of intrusion. The presented work aims at secure image transmission where a random encryption algorithm is used to encrypt different shares of stego image, which is created when a secret image and the cover image are embedded together, and produces shares using a secret sharing technique. At the receiving end, decryption of the encrypted shares are done using an artificial neural network and hence eliminating the need of key exchange before the transmission of data which is a prerequisite for most of the general encryption algorithm. Artificial neural network is used to provide high security of data and to produce distortion less decrypted images. The reversible process is used to reconstruct the secret image from the decrypted images. Peak signal to noise ratio, structure similarity and mean square error are used to analyze the quality of stego images. The simulation results show that the secret is reconstructed without loss and the time taken for encryption and decryption is very less.

In the process of ventricular premature beat (PVC) and normal sinus rhythm (NSR) identification base on electrocardiogram (ECG), there exists problems like negative effect from ECG rhythm and low recognition rate. This paper proposes the electrocardiogram PVC classification algorithm based on support vector machine (SVM) and wavelet algorithm. The algorithm uses the wavelet transform to analyze ECG beating model, which is not influenced by the change of ECG waveform. The two feature sets respectively compose of statistical parameters of the wavelet coefficients and the selected wavelet coefficients. PVC and NSR are analyzed by using SVM. The experimental results show that this method improves the recognition rate of ECG.

Multi-focus image fusion can fuse multiple source images with different focus settings into a single image that appears sharper. How to effectively and completely represent the source images is the key to multi-focus image fusion. A multi-component fusion method is proposed for multi-focus image fusion. The registered source images are decomposed into cartoon and texture components by cartoon-texture image decomposition. The salient features are selected from the cartoon and texture components respectively to form a composite feature space. The local features that represent the salient information of the source images are integrated to construct the fused image. According to the visual perception and objective evaluations on the fused images, the proposed method works better in extracting the focused regions and improving the fusion quality, compared with the other existing single-component fusion methods.

Graphical User Interfaces (GUI) is becoming increasingly important in the software field as it builds a friendly way between users and software through continuous interactions. A well-developed GUI is therefore an important factor of software quality. In particular, the reliability of GUIs is still on the way of development. Existing software reliability assessment techniques attempt to statistically describe the software testing process and predict the reliability of the system. However, those techniques are not suitable for GUI as quality of GUI is challenged by immense number of event interactions and complex structural profile etc. Furthermore, GUI has a wealth of information about GUI architecture, components, windows and their interactions with each other, which can be adopted to guide the testing process and establish confidence assessment of GUI. In this paper, a Bayesian network model of GUI reliability is introduced to discuss the reliability model topology and its issues encountered in the modeling process. A case is also presented to verify the validity of the model during the GUI reliability assessment process.

The spatio-temporal (ST) position information between local features plays an important role in action recognition task. To use the information, neighborhood-based features are built for describing local ST information around ST interest points. However, traditional methods of constructing neighborhood, such as sub-ST volumetric method and nearest-neighbor-based neighborhood method, ignore the orientation information of neighborhood. To make the neighborhood-based features more discriminative, we construct a novel, oriented neighborhood by imposing weights on the distance components. Specifically, in our scheme, firstly, local features are produced, and encoded by locality-constrained linear coding (LLC). Then, oriented neighborhoods are constructed by imposing weights on the distance components between features, and obtain single-scale oriented neighborhood features (SONFs). Next, multi-scale oriented neighborhood features (MONFs) are formed by concatenating SONFs. As a result, action video sequences are represented as a collection of MONFs. Finally, locality-constrained group sparse representation (LGSR) is used as classifier upon MONFs. Experimental results on the KTH and UCF Sports datasets show that our method achieves better performance than the competing local ST feature-based human action recognition methods.

According to the maximum coverage problem in wireless sensor networks, GA algorithm combined with the standard processing method can improve the network coverage based on, but there is a risk of falling into local optimum, and costs more computation time. On the condition of analysis and proof of the effectiveness by the employment of normalization processing to resolve coding redundancy of MCSDP, further details of the evaluate conditions are proposed, and new parent selection mechanism is introduced, which both are verified by compared experiments. The experiment results show that the optimization processing method proposed in this paper retains the characteristics of existing methods, and has better optimization performance and improves the network coverage rate as well as calculation speed, which verifies the effectiveness and superiority of the method proposed in this paper.