Earticle

This paper mainly studies a kind of different from the traditional signal processing theory, compressed sensing. The theory in the process of data transmission greatly save the space and the cost of storage and transmission is a new breakthrough in the data mining technology. Application of compressed sensing principle, the discrete tracking algorithm combined with matrix learning algorithm, to deal with the noise of original image. Through a large number of experiments show that the fusion algorithm is better than other similar algorithms in terms of denoising, is a better image denoising technique.

In this paper, a denoising algorithm and simulation experiments of algorithm based on wavelet transform and support vector machine (SVM) image is proposed, a new method is adopted in the selection of characteristic vector of support vector machine, based on training of support vector machine, the support vector machine model is used to distinguish between noise and the original image, to achieve the effect of denoising. The experimental results show that the method can well remove the noise, and can save some important details of images, compared with other denoising method based on wavelet transform, it has a good advantage.

In the process of image acquisition and transmission, noise is always contained inevitably. So it is necessary to image denoising processing to improve the quality of image. Generally speaking, each algorithm has some filtering and threshold parameters. Taking variety kinds of images into account, it is a key problem of how to set these parameters in denoising algorithms under different conditions to achieve better performance. There are many algorithms for the determination of the parameters, and each of them has its application field. Because the wavelet transform has good performance, therefore, it has been widely applied as a kind of signal and image processing tools. In this paper, wavelet transform is used in the image denoising, and the genetic algorithm is used to estimate the denoising results. Experimental results show the validity of the new algorithm.

In order to provide more comprehensive and effective information for cancer diagnosis and tumor treatment planning, it is necessary to fuse multi-modal medical images, such as CT/MRI image, CT/PET image, MRI/SPECT image and so on. In this paper, a multi-modal medical image fusion method based on Non-subsampled Shearlet Transform is proposed. Firstly, in this method, source images are decomposed into low-pass and high-pass subbands by NSST. Then, due to the characteristic features--large sparsity and strong contrast, the high-frequency and low-frequency coefficients of the images are fused by the different fusion rules. Finally, the image is reconstructed by the inverse non-sampled shearlet transform. In the method, the fusion rules are designed based on the regional energy and the average gradient; the image entropy, relative quality, average gradient, standard deviation and spatial frequency were used to evaluate the fusion results objectively. In the experiments, CT and MRI images are chosen to verify the method. Both the visual and the objective analysis show that the proposed method is better than the conventional Wavelet-based and non-subsampled Contourlet-based methods.

Model retrieval is widely applied to many fields including computer aided design, feature modeling and computer aided manufacturing. In order to retrieve a satisfied CAD model from a large model library, a new method to compute the similarity between two CAD models is presented in which a bipartite graph is used to match source faces with target faces. The number of edges in faces is extracted from CAD models. Based on the number of edges in faces, the similarity between source face and target face is computed. The maximum matching weight is calculated by KM algorithm. Then, the similarity between two CAD models is gotten. Several CAD models are given and their similarities are computed. Experimental results show that the method can evaluate the similarity between source CAD model and target CAD model efficiently.

Knowledge mapping will undoubtedly bring great convenience to application users for being behind the strong support of knowledge base. In this paper, we study how to discover the evolution of knowledge map in multi-languages. Our approach is uniquely designed to capture the rich topology of semantic items and to link the sub-graph to a global knowledge map. Instead of building a knowledge map start from scratch, we conceptually define semantic classes as a quantized unit of evolutionary link in sub-graph and discover new knowledge with multi-language dictionaries. Discovered new knowledge items are then connected to form an evolution knowledge map using a measure derived from the underlying semantic classes. We integrate these noisy items and entities into a unified probabilistic knowledge map using ideas from graph-based algorithm.

With the continuous development of computer science and technology, image processing and analysis gradually form the scientific system. Although history of image processing is not long, it attracts many researchers study on it. Digital media image widely exists in many fields, such as education, video, advertisement, and so on. Process digital media image is an important part of image processing. When analyze the digital media image, we want to extract the image part we care from the original image and then method for image segmentation is quite important. That is to say that the image segmentation will divide the image into a number of regions with specific and unique nature. How to keep the original characteristics of the digital media image is quite important in the image segmentation. In this paper, we propose a new algorithm for digital media image segmentation, and it is also can be used in the image processing. The algorithm is based on asynchronous particle swarm optimization algorithm to obtain the adaptive threshold; take the inertia factor into the algorithm, the optimal threshold has been acquired for the image segmentation. Compared with other particle swarm optimization algorithm, the algorithm has the advantages of stable, easy to converge to the optimal solution, and high segmentation speed.

Subdivision surface method is a series of iterative operation adopts a certain subdivision formula for an initial grid, obtains the smooth limits surface finally, and can dispose any arbitrary complex topology grid. At present most of the subdivision algorithm are 1-4 subdivisions and as the number of subdivision increase, the grid grow so too-rapid in the number of patch that it is difficult for the model after subdivision to deal with other things. We proposed an adaptive Doo-Sabin Mode subdivision algorithm to solve this problem, which take the average vector of the vertex and the angle between the intersecting surfaces of the vertex as a measurement criterion. This criterion is used to divide the surface, and then make local subdivision. In this way, when the times of subdivision are fewer (the demand of smoothness is not too high), the effect of subdivision has little difference, but efficiency of the algorithm can be greatly improved. Compared with the normal Doo-Sabin subdivision model, experimental results showed that adaptive Doo-Sabin subdivision algorithm can largely slow the growth speed of the amount of model data on the premise that guarantee the quality of surface.

A novel method for tracking underwater objects in three-dimensional space is proposed. Specifically, by a single camera this method has the ability to estimate the three-dimensional trajectory of the interest objects. Mathematically the general framework of this method is formulated as a probabilistic problem. In addition to the two-dimensional location in the image plane, the parameter respect to the distance between objects and the viewer (depth) is introduced into the motion state descriptor. The haze concentration in the image is utilized as the cue to estimate the depth. In order to match the objects existing in successive frames, color models of the reference and candidates are related by special consideration on the dynamic appearance variability caused by the underwater environments and the motion of objects. As a result, the trajectory exactly describing the three-dimensional motion trend of the interest objects is produced by the video filmed by a single camera. We present results over real underwater videos. Both single-object and multi-object tracking in three-dimensional space are achieved by using the proposed method.

Recently, photonic crystal fibers (PCFs) have attracted many researchers because of their unique properties, and design flexibility that can’t be realized by conventional fibers. One of the fruitful areas of research is the optical fiber hydrophone. In this paper, the finite element solver (FES), COMSOL multiphysics, is used to study and compare the response to acoustic pressure of a hollow-core photonic band gap fiber (HC-PBF), a solid-core photonic crystal fiber (SC-PCF), and a conventional single-mode fiber (SMF) for different acoustic pressures in the frequency range from 10 kHz to 50 kHz. The key structural factors affect the sensitivity to acoustic pressure (S) of the microstructured fibers are studied and a mathematical formula describes the relation of S and the dominant structural factor is proposed. Simulation results of the investigated optical fibers show that the normalized responsivity (NR) of the HC-1550, LMA-5, and SMF are -344 dB, -367.5 dB, and -366 dB, respectively. The proposed simulation results are in good agreement with published theoretical and experimentally measured results. The proposed results indicate the significance of the HC-PBFs in the future hydrophone systems and are useful for the design of microstrustured optical fibers for sensing applications.

A mathematical model of neurons network cells was studied. Firstly the study conducted numerical calculation by the C language program, and then simulated drawings with the grapher to analyze the model's complex dynamic conduct under different control parameters in the change interval of bifurcation parameter. Period-adding bifurcation and period doubling bifurcation must exist in the graph of ISI for different control parameters. As the external electric fields join in, the dynamic conduct of neuron model will change under corresponding parameters. The research results indicate external electric fields can change the discharge range of the model, which is helpful to explore the influence of external electric field on the excitability of biological neural system and pathogenic mechanism of neural system disease.

In medical Image analysis, the parallel segmentation is the core technology. As one of the classical methods, regional growth algorithms have some problems: it is hard to confirm the feed points automatically. To solve this defect, a new parallel segmentation algorithm with regional growth and support vector machine (SVM) is proposed. SVMs have a good result in segmentation (classification) but a non-ideal convergence rate which is the advantage of regional growth method. So that, combining them and the idea of the algorithm is: classify by SVM to search the seed points, segment by regional growth method. A curvature flow filter is also used in this algorithm to reduce the noise. The experiments are performed on a parallel environment based on torque. The results show that the algorithm is faster than conventional algorithms and the results are better.

A fiber-optic sensor is a sensor with high sensitivity, and signal detection technology is one of its key technologies. Phase generated carrier demodulation technology is a homodyne demodulation method widely used for interference fiber-optic sensors. Research reveals that the amplitude of the interference signal and the modulation depth will affect the result of demodulation. The distortion of the demodulated signal can be reduced by the elimination of the effects of the interference signal amplitude and modulation depth. A phase generated carrier demodulation method for interference fiber-optic sensors is proposed to detect weak ELF signals. Theoretical analysis and simulation have been implemented on the demodulation of weak ELF signals using low-level sub-mixer carrier. The results indicate that in the demodulation of weak ELF signals, DC drift occurs in the DCM approach, while the Arctangent approach does not have this problem. Then, an optical fiber sensing system for ELF signals is established based on optical fiber interferometer, indicating the Arctangent approach is able to demodulate the VLF signals accurately.

In order to effectively improve fusion quality, a novel multi-focus image fusion approach with sparse decomposition is proposed. The source images are decomposed into principal and sparse components by robust principal component analysis (RPCA) decomposition. A sliding window technique is applied to inhibiting blocking artifacts. The focused pixels of the source images are detected by using the salient features within the sliding window and integrated to construct the final fused image. Experimental results show that the proposed scheme can significantly inhibit the blocking artifacts compared to the other existing fusion methods in terms of some visual and objective evaluations.

A compression algorithm based on Empirical Mode Decomposition (EMD) is described in order to investigate the performance of EMD in biomedical signals, and especially in the case of electrocardiogram (ECG). The proposed algorithm is computationally simple to treat non-stationary and nonlinear data without pre- or post-processing. In order to evaluate the performance of the proposed compression algorithm, MIT-BIH arrhythmia database is applied, and the compress ratio (CR), percent root mean square difference (PRD), percent root mean square difference normalized (PRDN), root mean square (RMS), signal to noise ratio (SNR), and quality score (QS) values are obtained. When compared, good fidelity parameters are yielded with high CR as compared to wavelet transform (WT).

Maternal electrocardiogram (MECG) suppression is a critical procedure to estimate fetal heart rate (FHR). The impulse noises in maternal abdomen signals (ADS) include electromyographic interference (EMGI) and fetal R waves, which deteriorate FHR estimation in MECG suppression. In this paper, we introduce the M-estimation methods, a more robust methodology, to extract pure MECG in impulse noises. In order to evaluate the performance of M-estimation methods, two innovative quantitative performance metrics, namely loss degree of fetal R wave and sensitivity of initial value, are proposed. Numerical results show that: (1) FHR estimation using M-estimation methods is more accurate than least square method (LSM) in impulse noises; (2) the robust performance of Fair is better than that of Cauchy in low SNR; (3) the robust performance of Cauchy is superior to that of Fair in high SNR.

Finger-knuckle-print (FKP) pattern has been utilized in biometric recognition systems. This paper proposes a new FKP region of interest (ROI) extraction method based on two-stage center point detection. In our method, a center point preliminary detection is constructed to capture the center point initially. Then, an efficient center point positioned algorithm is presented to locate the center point more precisely in real time. Finally, we select the Hong Kong Polytechnic University (PolyU) database to verify the efficiency of the proposed method. The experimental results show that the proposed method can extract ROI not only accurately but also in real time.

This paper presents a robust and efficient method for vehicle detection in dynamic traffic environments. First, two adaptive vehicle hypothesis generation methods based on shadow and vehicle wave are presented, and then we assemble these two features into vehicle hypothesis. A hypothesis verification algorithm based on vehicle motion trajectory is proposed, the on-line hypothesis verification algorithm based on vehicle motion trajectory can not only reduce the false positive alarm caused by interferences, but also handle the problem that the classifiers generated in the off-line training phase is closely related to the diversity of positive and negative samples. Quantitative analysis on both public vehicle image datasets and real-time video presents a result of 85.58% detection rate with 4.13% false positive rate. And our algorithm could run as fast as 40ms/frame on PC platform.

Taking the high-resolution remote sensing image of industrial waste dump as the research object, this paper proposes a new image segmentation method based on marker-controlled watershed transform and region merging. The method gets the finial partition result by two-phrase segmentation on the pan-sharpened true color ALOS image. In the first phase, color gradient image of original image should be calculated and then it is modified by morphological impose minima, which should use markers extracted in two different ways. Lastly, preliminary segmentation result is obtained by watershed transform operates on the modified color gradient image. To solve the over-segmentation of industrial solid waste and other ground objects, region merging operation is performed according to the similarity measure criterion based on segmented objects’ color histogram Bhattacharyya coefficient in the second phrase, and then the final result is obtained. This method has been tested on the pan-sharpened ALOS image of 2.5 meters resolution in Shizuishan industrial zone, China. Experiment results demonstrate that this method is feasible and effective to segment the remote sensing industrial solid waste image.

Human action recognition has been a widely studied topic in the field of computer. However challenging problems exist for both local and global methods to classify human actions. Local methods usually ignore the structure information among local descriptors. Global methods generally have difficulties in occlusion and background clutter. To solve these problems, a novel combination representation called global Gist feature and local patch coding is proposed. Firstly, Gist feature captures spectrum information of actions in a global view, with spatial relationship among body parts. Secondly, Gist feature located in different grids of the action-centric region is divided into four patches according to the frequencies of action variance. Afterwards on the basis of traditional bag-of-words (BoW) model, a novel formation of local patch coding is adopted. Each patch is encoded independently and finally all the visual words are concatenated to represent high variability of human actions. By combining local patch coding, the proposed method not only solves the problem that global descriptors can not reliably identified actions in complex backgrounds, but also reduces the redundant features in a video. Experimental results performed on KTH and UCF sports dataset demonstrate that the proposed representation is effective for human action recognition.

This paper deals about image processing of log internal decays using MATLAB technique. Specifically, image enhancement, image edge detection and decay area extraction and calculation on the two-dimension image of log inner decays that were derived by using the non-destructive inspection technique of stress wave were conducted. The results showed that the testing image of stress wave can visually reflect wood inner decay, and the image resolution can also be improved by using the functions of image enhancement and edge detection in MATLAB software. The detection precision of stress wave is closely associated with the ratio of log internal decay area to tested wood cross-sectional area. When the ratio between wood actual inner decay area and tested log cross-sectional area increased from 4.77% to 45.52%, the relative error between tested decay area and actual decay area reduced from 16.42% to 7.39%. The study has provided an advanced method to judge the degree of wood inner decay for most forest practioners and researchers. Through the image processing and computing on the images derived by stress wave testing, the accuracy of logs internal decay discrimination can be significantly improved.

To the problem of face detection in color images, a novel method is proposed based on skin color segmentation and geometry features. Firstly, some common color models are analyzed, and a large amount of skin images are used to establish an YCbCr color model for region segmentation. Then, the morphological processing is executed on the binary image, and the facial regions filtering is conducted by adopting some geometry constraints such as Euler number, the ratio of width and height, centroid. Finally, the face region is located and labeled with a rectangle. Experimental results showed that the proposed method can equilibrate the unbalance between good accuracy and real-time performance to some extent, and is effective in face location.

In order to locate the license plate in noisy background or various illumination conditions, a license plate location model based on visual attention is proposed, which imitates the biological mechanism of human visual. Four major visual sensitive characteristics of Chinese license plate are applied. First, the features of intensity, color and texture are extracted and integrated to structure the saliency map; second, the obtaining the optimal threshold based on the shape feature for foreground segmentation; at last, support vector machine is applied to eliminate the false plates in the image. Experimental results show that the proposed model is adaptive to various illumination conditions, complicated background, and resistant to noise.

Background subtraction is important process in many automatic video content analysis applications. extracting foreground moving objects from video sequences is a crucial step and also a hot topic in computer vision and image processing . This paper presents a new approach in image processing for detection crack in video. This method involves two main steps: First, we use based on Visual Background Extractor (Vibe) to detection moving crack in video. Second steps: Using a suitable threshold in a binary image and classifies all pixels two groups’ background and foreground. And use filter area and changes the area if less than the specific number to back; and using filtering to elimination of residual noise. This paper describes a method for detection crack in video we use Image Processing Techniques. The advantage of this method is clearly and accurate detection of cracks in video. Experimental work shows that our method is improved relatively to the other widely used techniques.

This paper presents a methodology to recognize certain crop fields’ images using texture, color and combination of both types of features. In this work, we have considered eight varieties of crop images, namely, Brinjal, Cotton, Groundnut, Paddy, Soyabean, Sugarcane and Sunflower. Texture features using GLCM and color features using HSV are deployed. Artificial Neural Network (ANN) is used for recognition. Considering only as feature, classification accuracies of 63.75%, 66.25% and 84.375% are obtained using texture, color and their combination respectively. The work is helpful in the area of agriculture for early detection and prevention of diseases.

This paper proposed brightness preserving dynamic fuzzy histogram equalization using triangular membership function which is the modified technique of histogram equalization. This modified technique, called Brightness Preserving Dynamic Fuzzy Histogram Equalization (BPDFHE), uses fuzzy statistics of digital images for their representation and processing in the fuzzy area which enables the technique to handle the approximation of gray level values in a better way for better presentation. This algorithm enhances image contrast as well as conserves the brightness very well. Some images are not available to excellent quality, so proposed Fuzzy algorithm can be used for image enhancement to improve the quality of the image.

Conventional attitude/orientation estimating filters are generally complex demanding excessive computational burden. This instigates the requirement for a computationally simple yet sufficiently precise algorithm for applications where computational complexity is of prime import. A relatively simple, robust and equally efficient technique in this regard is the development of complementary filters. Gradient Descent based Complementary Algorithm (GDCA) and Explicit Complementary Clgorithm (ECA) are the latest advancement in complementary filters applicable to low cost, low power MEMS based inertial measurement units (IMUs) employing quaternion. These fixed gain estimators employ gyroscope triad for high frequency estimation and accelerometer triad for low frequency attitude estimation. This paper appraises the performance of GDCA and ECA. Both simulation and experimental results are presented for comparative analysis. Simulated data was generated in MATLAB for known orientation in term of Euler angles to validate the filters performance whereas for practical implementation of different scenarios, MEMS based MPU6050 IMU was employed. As only IMU is employed without aided sensory system, the mandate of this research is limited to attitude estimation in terms of Euler roll and pitch angles. Roll of the adjustable filter gains is also assessed for a range of values.