Applied Mechanics and Materials Vols. 433-435

Abstract: Automatic focusing is one of the key technology of robot vision and digital video-systems, while play an important role in determining the quality of image. The performance of focusing depends on whether the evaluation function has unbiasedness, unimodality and noise resistance. This paper proposes a new evaluation function algorithm by improving image clarity-evaluation function of the traditional neighborhood difference operator. Compared with the existing algorithm, the results of experiments demonstrated the new algorithm has a good sensitivity, timeliness, good anti-noise ability and stability during the automatic focusing process.

Abstract: A method based on moments invariant features and cascade AdaBoost classifier for insulator recognition is put forward to solve the problem of poor performance of insulator recognition. At first, the insulator image is preprocessed by median filtering, dilating, eroding and Otsu thresholding. Then, for the better extraction of moments invariant features, the preprocessed insulator image is tilted correctly based on PCA (Principal Component Analysis). Next, the moments invariant features are extracted and chosen to compose complex classifier in the process of training AdaBoost. Finally, the complex AdaBoost classifiers are combined in a cascade method for insulator recognition. The results of experiments demonstrate that the proposed method can recognize the insulator from complex background in the mountainous area, and it has better robustness, accuracy and validity.

Abstract: Mutual information stems from communication theory, which is commonly used as similarity measure in the field of medical image registration. This approach works directly with image data; no pre-processing or segmentation is required. But calculating the mutual information of images needs a large amount of computation, which in some respect restricts its application. In this paper, by doing some processing on the reference image before the registration, we changed the way of calculating the mutual information to reduce the computation. The result of the experiments shows that the accuracy of registration does not change significantly, whereas the time of calculating the mutual information is decreased significantly.

Abstract: Image quality assessment is an important issue in the area of image processing, and the no-reference image quality assessment tries to evaluate the quality of image without the reference image. The present no-reference image quality assessment approach can not predict the quality score accurately. This paper proposes a new image quality assessment approach based on two-dimensional discrete fractional Fourier transform (FRFT). After the image is processed by two dimensional discrete fourier transform, the histogram of FRFT coefficients in different order are modeled by generalized Gaussian distribution (GGD). The parameters of GGD are estimated and the feature vector is formed by parameters of GGD. After that, the image is classified into five distortion type by the trained support vector machine. At last, the quality score is predicted by the trained support vector regression machine. The experiment results show that the performance of proposed method is better than the traditional method.

Abstract: As to the speech signal processing problems under the complex voice environment, the characters of speech harmonic and the structure of voiced harmonic are discussed in this paper. In the third dimension frequency domain, quadratic Fourier transform algorithm based on logarithmic amplitude-frequency characteristics is used to propose the concept of the third dimension spectral harmonic ratio" by the behavior of the quasi-sinusoidal characteristics of the speech short-time Fourier spectrum slicing, which is cosider to be an important basis to discriminate speech activity detection. The concept of third dimension spectral harmonic ratio maks the speech signal as a special signal, and separate from the other noise signal completely. For voice noise outside the speech segment, it is no longer need to discuss its own characteristics and without the need for pre-processing to shield the noise accurately, which bring new ideas for the speech signal detection in the complex noise environment.

Abstract: In order to remove salt-and-pepper noise and Gaussian noise in image, a novel filtering algorithm is proposed in this paper. The novel algorithm can preserve image edge details as much as possible. Firstly, five-median-binary code (FMBC) is proposed and used to describe local edge type of image. Secondly, median filter algorithm is improved to remove salt-and-pepper noise by using FMBC. Then, local enhanced bilateral filter with FMBC and a new type of exponential weighting function is used to remove Gaussian noise. Simulation results show that the algorithm proposed in this paper is very effective not only in filtering mixed noise but also in preserving edge details.

Abstract: In the field of military communications, the SNR of speech signal is low because of the additive noise,. The two-level Wiener filter is adopted to analysis the noise characteristics. The DSP system architecture solutions are given out in this paper.The algorithms of the PSD mean module, Mel scale filter bank module, voice activity detection module and gain control module are introduced in detail. Finally, the application results are given out, which indicate that two Wiener filter can achieve better results, in short stationary noise environments when face the low SNR.

Abstract: Detection model of knotty illicit image is designed to detect special erotic image, such as nude chat image and advertising prostitution image, which are difficulty identified by conventional algorithm. Text information and skin-color information are detected in the image detection windows set according to the characteristic of images. Experimental results demonstrate that the model can obtain good results in detecting special erotic image.

Abstract: An image enhancement method in mixed space is proposed in this paper, it combines the Laplace operator and the Sobel operator, gives full play the advantages of the two algorithms. It shows that the image enhancement effect of the mixed spatial method is better than the Laplace method and gradient method, through the enhancement experiments to the same image in the three ways, comparing and analyzing the results of their treatment.

Abstract: In order to alleviate the effect of illumination variations and improve the face recognition rate, this paper proposes a novel non-statistics based face representation method, which is called Center-Symmetric Local Nonsubsampled Contourlet Transform Binary Pattern Histogram Sequence (CS-LNBPHS). This method first applies NSCT to decompose a face image, and obtains NSCT coefficients in different scales and various orientations. Then, CS-LBP operator is used to get CS-LBP feature maps from NSCT coefficients. After that, feature maps are respectively divided into several blocks, the concatenated histogram, which are calculated over each block, are used as the face features. Experimental results on YaleB, ORL face databases show the validity of the proposed approach especially for illumination, face expression and position.