Applied Mechanics and Materials Vols. 568-570

Abstract: This paper includes two parts. Firstly, we introduce the pretreatment process of fish photos. This is a very important part for fining treatment,because the quality of the image taken from fishpond would be greatly affected by luminance and complex background. Finally we give a description for the thinning algorithm about the shape of fish. The pretreated fish photos could be processed by thinning algorithm to make each fish similar to string. It’s very easy to judge the dynamic behavior of fish. In the fish farms, you should continue to feed if the lines assume radial shape. You should stop feeding if the lines present us with random shape.

Abstract: The application of digital technology on the protection of intangible cultural heritage is a major topic of research in recent years. The motion capture technology of protection will gradually replace the traditional recording methods such as texts, pictures and videos. It is valuable to build a high-fidelity, high-modular and low-cost digital platform for choreographic data collection and extended application. This paper studies the intangible cultural heritage of Quanzhou breast-clapping dance, one of the most famous choreographic intangible cultural heritages from China with standard optical motion capture method. The data are acquiring and processing after the dance motion capture, we binds the motion data and three-dimensional model using Motion Builder and build digital demonstration platform base on an OGRE engine to display the movements. The viewer can view at any angle and distance. The system can be easily applied in motion intangible cultural heritages protection project. Furthermore, the system can be provided versatile motion data for additional use.

Abstract: This paper presents a projection on the basis of the true 3D imaging display system and analyzes construction for the voxel. We proposed a new method based on the theory of high-speed projector - Three DMD/OLED coaxial optical path display. We did some experiments about the clock distribution for the coaxial optical path. The innovation improves frame rate and operating stability for the display device, reduces the high frame rate on one single device. We also found a new approach - local state-space model - Fornasini-Marchesini for constructing the volume pixels. The mechanism is simple to build and low price; the optical field can be displayed in the whole above x-y plane except the zenith point.

Abstract: For the identification complexity of rock microstructure, based on numerical analysis of rock section images, an automatic rock texture classification method and identification system is proposed in this paper. Digital grey image processing of rock thin section is used for features extraction, the features are then as inputs to the neural network model, the model output is the rock microstructure classification. 100 pieces of rock section images from Sulige region in Changqing Oilfield are used for the experiment; the whole dataset is randomly divided into 70 images for training datasets, 15 images for validation datasets and 15 images for testing datasets. It is shown that the correct classification rate for automatic identification of rock microstructure is about 93.3%. Therefore, the proposed method for solving geological problem is effective and can get a good identification performance for rock microstructure classification quickly and accurately.

Abstract: The paper studies the method of behavior recognition. The standing, sitting, squatting postures are recognized in the PC and the Davinci processor. The paper achieves the method that detects the moving regions using codebook method. After the moving regions being smoothed and chosen, then extract the features of regions. Last we train the examples using the K near algorithm and recognize the posture in the video in real time.

Abstract: In this paper the problem of accurate edge detection for pulp fibers images is discussed. The images are acquired by the smart CCD camera. Subpixel edge detection is considered that applied in the system to improve the measurement accuracy of pulp fibers. The location of a precise edge is the basis of kink index, kink Angle, curl index and other measurement parameters for the pulp fibers. A new method for subpixel edge detection of the pulp fibers is introduced. Canny operator is used to get the pixel level edge. Each edge point determines five pixels in the gradient direction. Then the least squares fitting is used to locate its sub-pixel edge. The measurement results show that the proposed subpixel edge detection method is valid to improve measurement accuracy for real time measuring the pulp fiber parameters.

Abstract: On the foundation of analyzing in an all-round way to the current application situations and development trends with image acquisition and treatment technology on domestic and international, the paper designed one system that aim at the problems including high-speed rate of analog-to-digital conversion, heavy amount data and complicated arithmetic and so on that existing in high-speed image acquisition board and realize the high-speed acquisition and processing of image. The key part of the system is DSP and programmable arrays device FPGA, The main circuit including the standard video signal separation circuit, A/D converter circuit, phase locking multiple frequency circuits and the DSP, such as power supply circuit, etc. The circuit is simple, dependability, and have certain common ability.

Abstract: To improve the recognition rate and resolve the view problem in gait analysis, a view-invariant method is presented. This method extracts the mean shape of the head and shoulder models in a gait cycle by using the Procrustes Shape Analysis (PSA) algorithm. Then, the mean shape is converted from 2-D space to 1-D space by unfolding the 2D mean shape from the top point of the head. Furthermore, the piecewise approximation (PA) of the unfolded mean shape is segmented corresponding to the average distance value. The obtained PA is adopted for estimating views and recognizing the gait. Finally, the presented method is conducted on the CASIA B database. The results illustrate that the PA is a powerful discriminative feature for view estimation and gait recognition, and the performance has been improved when there is view variation.

Abstract: This paper presents a wavelet energy map based on an active contour model for medical image segmentation and bias correction in a variational level set framework. In our model the wavelet transform amplifies the faint dissimilarities between regions, and we model the distribution of intensity belonging to each tissue as a Gaussian distribution with spatially varying mean and variance. In addition, we modify the intensity mean as the product of the bias field and true image. Tissue segmentation and bias correction are simultaneously achieved via a level set evolution process. Experiments on images of various modalities demonstrated the superior performance of the proposed approach.

Abstract: In this paper, we present a novel no-reference blur metric for images. The blur metric is based on analyzing image features include the mean value of phase congruency image, the entropy of phase congruency image and the distorted image, and the gradient of the distorted image. The new index does NOT need any information from reference image, and image quality estimation is accomplished by simple functional relationship between those features. Our experimental results show that the new index outperforms existing popular no-reference blurriness metric and full reference PSNR on LIVE Gaussian blurred database and IVC blurring images.