Applied Mechanics and Materials Vols. 321-324

Abstract: After background modeling and segmenting of moving object for surveillance video, this paper firstly presented a noninteractive matting algorithm of video moving object based on GrabCut. These matted moving objects then were placed in a background image on the condition of nonoverlapping arrangement, so a frame could be obtained with several moving objects placed in a background image. Finally, a series of these frame images could be achieved in timeline and a single camera surveillance video synopsis could be formed. The experimental results show that this video synopsis has the features of conciseness and readable concentrated form and the efficiency of browsing and retrieval can be improved.

Abstract: The support vector machine (SVM) has been shown to be an efficient approach for a variety of classification problems. It has also been widely used in target identification and tracking, motion analysis, image segmentation technology. Traditional detection methods mostly exist pseudo-edge and poor anti-noise capability. Under these circumstances, developing an efficient method is necessary. In this paper, we propose a new detection algorithm based on FSVM, the main idea is to train classified sample and give all training data a degree of membership, increase punishment to the wrong sub-sample. Then training and testing the FSVM classification model. Finally, extract edge of the image by using FSVM classification model. Experimental results show that the new algorithm can detect a clear image edge and have a good anti-noise nature.

Abstract: In this paper, a new image segmentation approach is proposed based on the morphological watershed transform and an improved balloon snake algorithm for reducing the incorrect segmenting rate that caused by uneven dyeing of Cerebrospinal Fluid (CSF) cells. Firstly, a marker-controlled watershed pre-segmentation was applied to resolve the over-segmentation problem of the watershed and eliminate the approximate contour of CSF cells. Secondly, due to the weak boundary of some cytoplasm, we performed a coarse segmentation by using an improved balloon snake to dilate the watershed result. Finally, a fine segmentation was performed to obtain the ultimate cytoplasm contours. The results show that the new method is more accurate and robust to weak boundary.

Abstract: At first, in the paper, through the preprocessing of face image by the histogram equalization, enhance the image contrast. Then this paper uses Gabor wavelet transform, by choosing different size and direction to facial features extraction and Principal Component Analysis method for dimension reduction. Finally, input the face feature into the Radial Basis Function neural network for identification. Based on RBF neural network is trained and weights optimization, the efficiency of identification had been improved.

Abstract: In order to improve the performance of the banknote classification, new banknote image feature extraction method is proposed in this paper. The contourlet transform is applied to the original banknote image which is obtained by image contact sensor.The statistical characteristics of transformed image in the contourlet domain are analyzed. The statistical characteristics which can perfectly reflect the banknote image texture information are used as feature vector for banknote classification. The experimental results show that the proposed method can obtain higher recognition compared with other conventional banknote image feature extraction methods.

Abstract: The paper illustrates the background of boutique open video class in China, and the differences between the boutique open video class and the boutique shared class. It also discusses the domestic study of the strategy of forming the online open courses initiated by foreign distinguished universities in recent years. By the comparison of the strategy used at home and abroad in terms of equipment, shooting techniques, and multimedia technology utilization, it draws the revelation of producing boutique video open classes.

Abstract: Based on the facts in medical images that there are many empty voxels and the near scene is more sensitive to human vision, our accelerating rendering algorithm divides the rendering process into near scene and far scene parts. An accelerated ray-casting method is adapted for rendering the near scene and a fast splatting method is adapted for rendering the far scene. The inherent coherence existed in adjacent frames is used to accelerate the navigation speed. A part of far scene of the next frame may be simply rendered by view transforming the part of far scene of the current frame.

Abstract: A higher intelligibility speech-enhancement algorithm based on subspace is proposed. The majority existing speech-enhancement algorithms cannot effectively improve enhanced speech intelligibility. One important reason is that they only use Minimum Mean Square Error (MMSE) to constrain speech distortion but ignore that speech distortion region differences have a significant effect on intelligibility. A priori Signal Noise Ratio (SNR) and gain matrix were used to determine the distortion region. Then the gain matrix was modified to constrain the magnitude spectrum of the amplification distortion in excess of 6.02 dB which damages intelligibility much. Both objective evaluation and subjective audition show that the proposed algorithm does improve the enhanced speech intelligibility.

Abstract: The edge detection is the important role in the image disposal. Traditional methods had some limitations more or less in practical applications such as pseudo-edge or need setting parameters by manual.Now, proposed a method can solve these problems in this paper. The histogram of gradient effective features was selected to composite the feature space, and during the process of classifier training, combined with AdaBoost and decision tree algorithm to improve the classification accuracy. Finally, the application of the algorithm proposed to image of Lena edge detection and comparative experimental show that the algorithm has better self-adaptability and good edge can be detected through this new algorithm.

Abstract: Three dimensional modelingis an important issue in the field of computer vision and computer graphics.And because of its universality and usability, human face is one of the mainfocuses of three dimensional modeling. At the same time, the wide spread of mobile internet makes people’s demand of mobile-working and mobile-entertainment become larger and larger.Based on this background, this paper designs and implements a mobile environment compatible 3Dface modeling system based on client-serverarchitecture, and proposes a solution to handle the useless and redundantface data during the pre-processing period of face modeling based on the frames’ continuity assumption.