Advanced Materials Research Vols. 718-720

Abstract: In order to improve the robustness and real time performance of SURF based image matching algorithms, a new descriptor is proposed. We compute the new descriptor in a rectangle local region (the side set to 20s). Firstly, the local region is divided into 8 equal triangle subregion. Secondly, local region location grid is rotated to align its dominate orientation to a canonical direction. The keypoint dominate orientation and its orthogonalorientation is defined as the x and y directions of the descriptors local coordinate system.Thirdly, compute the Haar wavelet response in x and y directions within the keypoint local region. In order to reduce the boundary effect and outer noise, Haar wavelet response in the same Grid of different triangle is both assigned to each triangle in different weight, and then a gaussian weighting function is used. Compute the histogram of Haar wavelet response and absolute Haar wavelet response, so each triangle subregion constitutes a vector with 4 dimensions. Finally, a descriptor with 32 dimensions is constituted and the descriptor is normalized to achieve illumination invariance. The experimental results show that the performance of the new descriptor is even better than SURF descriptor.

Abstract: A marginal clone algorithm is proposed in order to make edge detection of retinal vessels image have better edge continuity and less detection points. It is applied in edge detection of retinal vessels with a combination of edge connectivity and noise removal algorithm based on analyzing image edge cloning theory and algorithms. Simulation results show that the edge image detected by the algorithm has solved problems of traditional edge discontinuities and too much noise. Moreover, it has better edge recognition performance.

Abstract: Minimizing the nuclear norm is recently considered as the convex relaxation of the rank minimization problem and arises in many applications as Netflix challenge. A closest nonconvex relaxation - Schatten norm minimization has been proposed to replace the NP hard rank minimization. In this paper, an algorithm based on Majorization Minimization has be proposed to solve Schatten norm minimization. The numerical experiments show that Schatten norm with recovers low rank matrix from fewer measurements than nuclear norm minimization. The numerical results also indicate that our algorithm give a more accurate reconstruction

Abstract: Symbolic execution is a promising approach for software analyzing and testing, but it still suffers from scalability issues, in which a significant challenge is how to handle loop caused path explosion. This paper proposes a new approach to mitigate the scalability problem brought by loops in symbolic execution. By reusing the previously calculated results captured from satisfiability verification process, we implement a fast loop boundary coverage test strategy which can eliminate a large number of redundant paths and unnecessary solver invocations. Experimental results on Symbolic (Java) PathFinder show that our strategy can greatly improve the efficiency of testing for programs having loops.

Abstract: This paper proposed a new solution to the situation that the moving objects detection in natural environment might be disturbed by some natural conditions, such as wind, light, etc. Firstly, the background image was modeled with Gaussian mixture models, to eliminate the interference of some slight and cyclical movements caused by wind, like branches swinging. Secondly, the foreground was segmented through background subtraction, and the background model was updated to adapt to the gradual change of illumination. Finally, the shadows of moving objects were detected and removed in HSV color model, meaning that the detection was completed. The experiment verified that the solution above can effectively realize the preliminary detection of moving objects in natural environment.

Abstract: The number of projection affects the reconstruction quality in computed tomography system. Larger projection number will lead to better reconstruction quality. On the other hand, less projection number can reduce the radiation dose, save time and keep the patients comfortable. The optimal number of projection compromises both the mathematical accuracy and practical requirement. The purpose of this paper is to find the relationship between the projection number and the noise properties of reconstructed cross-section. It can provide references for researchers on CT application. A micro-CT system is utilized to validate the relationship between projection number and reconstruction quality in fan beam geometry. Full-scan and short-scan method are applied to reconstruct the specimen. Compared with full-scan method, short-scan can reduce radiation dose with the same step angle, and it can achieve better results with the same projection number when the sample data is sufficient.

Abstract: Classical mean shift tracking algorithm doesnt show good performance when the tracked objects move fast, change in size or pose. This paper proposes an improved mean shift method used for vehicle tracking. Firstly, a position prediction model based on second order auto-regression process is used to find the initial position of mean shift iteration, reduce times of iteration and enhance the tracking accuracy. Secondly, we employ a position search method based on the weight image to improve the tracking result when the result of basic mean shift tracking is not good. The proposed algorithm is tested in a real traffic video to track a vehicle changing in size and pose with more accurate result than basic mean shift tracking algorithm.

Abstract: Moving target tracking is a hot research spot of computer vision and applied in various fields. In this paper, a new tracking method base on sparse optical flow is put forward. In this method, targets are tracked through calculating the movements of Harris corner points, rather than the movements of all pixel points. Experiments results show that the tracking effect of this new method is pretty good. Tracking accuracy can reach more than 80% in most experimental conditions. And according to other peoples research production, experiments based on dense optical flow are done to compare with the new method proposed in this paper. The comparison results show that the new method has high calculation efficiency. This indicates that the method has feasibility and practical value.

Abstract: Mean of neighbor pixels is usually computed in image process, its computation cost is very high, which lead to the executing speed. A fast computation strategy is applied to get the mean of neighbor pixels. In our approach, the means of neighbor pixels located in fist column and line are computed, after that the iterative solution methods is adopted to compute neighbor means of pixels located in neighbor pixels of current pixel. The proposed approach do not change the filter result, but which can accelerate the running speed, especially for large image, finally, we used several image to test the approach, the experiment results indicate that the approach do not change the filtering premise and effectively reduce the computational complexity.

Abstract: A method to extract contour of wheat pests based on LCV model is proposed in this paper. This method uses calculation of the weighted average gray value of the partial window function to replace the global mean value, and added the constraint level set function as a the energy term of signed distance function to avoid the re-initialization of the level set function. This method can be extracted the contour of wheat pests effectively, and thus it improves the efficiency of the extraction of wheat pests shape feature. The effectiveness of this method is verified by MATLAB simulation experiments.