Search

In this paper, a new efficient control algorithm is proposed by coupling differential equations. Our unified model is natural for posing and solving hybrid analysis and controlling problems. We handle the famous asynchronous arbitrary problem in continuous and hybrid settings and analyze a hybrid digital signal processing that logically switches between two conventional controllers; meanwhile, global asymptotic stability for a realistic set of cases is proved. We systematize the notion of a hybrid digital signal processing by a hybrid controller using an optimal control framework. Theoretical results lead us to algorithms for synthesizing such hybrid controllers based on a generalized Bellman equation, impulse control algorithms, and linear programming.

3D object recognition has attracted considerable research in computer vision and computer graphics. In this paper, we draw attentions from neurophysiological research that line drawings trigger a neural response similar to natural color images, and propose a line-drawing-based 3D object recognition method. The contribution of the proposed method includes a feature defined for line drawings and a similarity metric for object recognition. Experimental results on McGill 3D shape benchmark show that the proposed method has the best performance when compared to five classic 3D object recognition methods.

In this paper, a novel soft computing modeling of adaptive neuro-fuzzy inference system in longitudinal vehicle control and warning systems is investigated. Fuzzy inference system on the outputs of neuro-fuzzy classifiers is proposed, making decision of whether the current activity is normal or intrusive. Finally, a genetic algorithm which optimizes the structure of the fuzzy decision engine is constructed. A simulation is undertaken to performance of the proposed method and results are shown promising.

In this paper a novel cross correlation technique is proposed for shape matching between two similar objects. The proposed technique can not only evaluate the similarity between any two objects, but also has two distinct advantages compared to previous work: (1) the deformed articulated objects such as human being with different poses, can be matched very well; (2) the local feature extraction and correspondence can be established at the same time. The basic tool we used is the shape profile driven from the curvature map of the object profile. The cross correlation technique is applied to the shape profile of the two objects to evaluate their similarity. Filtering scheme is used to enhance the quality of both shape matching and extracted features. The invariant property, the robustness and the efficiency of the shape profile in shape matching and feature extraction are discussed.

An important surface modeling problem in CAD is to connect two disjoint B-spline patches with the second-order geometric continuity. In this paper we present a study to solve this problem based on the surface extension algorithm [Computer-Aided Design 2002; 34:415- 419]. Nice properties of this extension algorithm are exploited in depth and thus make our solution very simple and efficient. Various practical examples are presented to demonstrate the usefulness and efficiency of our presented solution.

In this paper a computational model called dynamic medial axis ($\mathcal{DMA}$) is proposed to describe the internal evolution of planar shapes. To define the $\mathcal{DMA}$, a symbolic representation called matching list is proposed to depict the topological structure of the medial axis. As shown in this paper with provable properties, the $\mathcal{DMA}$ exhibits an interesting dynamic skeleton structure for planar shapes. Finally an important application of the proposed $\mathcal{DMA}$ — computing the medial axis of multiply-connected planar shapes with curved boundaries — is presented.

In this paper we address practical implementation issues of generating a progressive representation of point-sampled geometry. Fast and stable algorithms are proposed to efficiently implement a progressive version of the modified RBF Shepard's method. Comparisons with several well-known methods are presented, showing that the proposed algorithms can achieve good balance between geometric error reduction and time efficiency.