|
Recognition of 3-D objects based on Markov random field models
HUANG Ying, DING Xiao-qing, WANG Sheng-jin
Front. Electr. Electron. Eng.. 2006, 1 (2): 125-129.
https://doi.org/10.1007/s11460-006-0003-6
The recognition of 3-D objects is quite a difficult task for computer vision systems. This paper presents a new object framework, which utilizes densely sampled grids with different resolutions to represent the local information of the input image. A Markov random field model is then created to model the geometric distribution of the object key nodes. Flexible matching, which aims to find the accurate correspondence map between the key points of two images, is performed by combining the local similarities and the geometric relations together using the highest confidence first method. Afterwards, a global similarity is calculated for object recognition. Experimental results on Coil-100 object database, which consists of 7 200 images of 100 objects, are presented. When the numbers of templates vary from 4, 8, 18 to 36 for each object, and the remaining images compose the test sets, the object recognition rates are 95.75 %, 99.30 %, 100.0 % and 100.0 %, respectively. The excellent recognition performance is much better than those of the other cited references, which indicates that our approach is well-suited for appearance-based object recognition.
Related Articles |
Metrics
|
|
Edge detection based on directional space
YUAN Wei-qi, LI De-sheng
Front. Electr. Electron. Eng.. 2006, 1 (2): 135-140.
https://doi.org/10.1007/s11460-006-0014-3
A new method for edge detection based on directional space is proposed. The principle is that: firstly, the directional differential space is set up in which the ridge edge pixels and valley edge pixels are abstracted with the help of the method of logical judgments along the direction of differential function, forming a directional roof edge map; secondly, step edge pixels are abstracted between the neighboring directional ridge edge and directional valley edge along the direction of differential function; finally, the ridge edge map, valley edge map and step edge map gained along different directions are combined into corresponding ridge edge map, valley edge map and step edge map. This method is different from classical algorithms in which the gray differential values of the mutual vertical direction are combined into one gradient value. The experiment of edge detection is made for the images of nature scenery, human body and accumulative raw material, whose result is compared with the one of classical algorithms and showing the robustness of the proposed method.
Related Articles |
Metrics
|
|
Active contour model based on force field analysis
HOU Zhi-qiang, HAN Chong-zhao
Front. Electr. Electron. Eng.. 2006, 1 (2): 153-159.
https://doi.org/10.1007/s11460-006-0026-z
The traditional snake initial contour should be close to the true boundary of an object of interest in an image; otherwise, an incorrect result will be obtained. Next, active contours have difficulties progressing into boundary concavities. Moreover, the traditional snake as well as almost all of its improved methods can be easily obtained from the local minimum because snake models are nonconvex. An active contour model based on force field analysis (FFA), namely, FFA snake model, is presented in this paper. Based on analyzing force distribution rules of the distance potential force field, a standard is introduced here to distinguish the false one from contour points. The result is not considered as the final solution when the snake energy is minimal. Furthermore, estimation and calculation should be made according to the established standard; only then can the result be considered final. Thus, the snake is prevented from running into the local minimum. The simulation results show that the FFA snake model has a large capture range and can move a snake into the boundary concavities, and that it is able to obtain the object of interest s contour precisely. Compared with the gradient vector flow snake, this new model has a low computational cost.
Related Articles |
Metrics
|
|
Parallel mining and application of fuzzy association rules
LU Jian-jiang, ZOU Xiao-feng, XU Bao-wen, KANG Da-zhou, LI Yan-hui, ZHOU Jin
Front. Electr. Electron. Eng.. 2006, 1 (2): 177-182.
https://doi.org/10.1007/s11460-006-0024-1
Quantitative attributes are partitioned into several fuzzy sets by using fuzzy c-means algorithm. Fuzzy c-means algorithm can embody the actual distribution of the data, and fuzzy sets can soften the partition boundary. Then, we improve the search technology of apriori algorithm and present the algorithm for mining fuzzy association rules. As the database size becomes larger and larger, a better way is to mine fuzzy association rules in parallel. In the parallel mining algorithm, quantitative attributes are partitioned into several fuzzy sets by using parallel fuzzy c-means algorithm. Boolean parallel algorithm is improved to discover frequent fuzzy attribute set, and the fuzzy association rules with at least a minimum confidence are generated on all processors. The experiment results implemented on the distributed linked PC/workstation show that the parallel mining algorithm has fine scaleup, sizeup and speedup. Last, we discuss the application of fuzzy association rules in the classification. The example shows that the accuracy of classification systems of the fuzzy association rules is better than that of the two popular classification methods: C4.5 and CBA.
Related Articles |
Metrics
|
|
Modular design of adaptive robust controller for strict-feedback stochastic nonlinear systems
WANG Jun, XI Hong-sheng, JI Hai-bo, KANG Yu
Front. Electr. Electron. Eng.. 2006, 1 (2): 216-222.
https://doi.org/10.1007/s11460-006-0009-0
A modular approach of the estimation-based design in adaptive linear control systems has been extended to the adaptive robust control of strict-feedback stochastic nonlinear systems with additive standard Wiener noises and constant unknown parameters. By using Itô s differentiation rule, nonlinear damping and adaptive Backstepping procedure, the input-to-state stable controller of global stabilization in probability is developed, which guarantees that system states are bounded and the system has a robust stabilization. According to Swapping technique, we develop two filters and convert dynamic parametric models into static ones to which the gradient update law is designed. Transient performance of the system is estimated by the norm of error. Results of simulation show the effectiveness of the control algorithms. The modular design, which has a concise hierarchy, is more flexible and versatile than a Lyapunov-based algorithm.
Related Articles |
Metrics
|
|
Hybrid Method of combined difference and spectrum for time-domain maxwell’s equations
QI Hong-xing, CHEN Shu-de, QIAO Deng-jiang, PANG Xiao-feng
Front. Electr. Electron. Eng.. 2006, 1 (2): 248-250.
https://doi.org/10.1007/s11460-006-0023-2
For the finite-difference time domain (FDTD) method, the electromagnetic scattering problem, which requires the characteristic structure size to be much smaller than the wavelength of the exciting source, is still a challenge. To circumvent this difficulty, this paper presents a novel hybrid numerical technique of combined difference and spectrum for time-domain Maxwell s equations. With periodical continuation of each time-dependent quantity in Maxwell s equations, the solutions before and after the continuation remain consistent in the first period, which results in the conversion of the continuous spectrum problem to a discrete one. The discrete spectrum of the field after continuation is obtained from difference methods for Maxwell s curl equations in frequency-domain, and the time domain solution of the original problem is derived from their inverse Fourier transform. Due to its unconditional stability, the proposed scheme excels FDTD in resolving the aforementioned problems. In addition, this method can simulate dispersive media whose electric susceptibility cannot be expressed with Debye or Lorentz types of models. In dealing with boundary conditions, it can utilize the perfectly matched layer (PML) without extra codes. Numerical experiments demonstrate its effectiveness, easy implementation and high precision.
Related Articles |
Metrics
|
23 articles
|