Catálogo de publicaciones - libros

3D video mosaics created by 3D image mosaicking, our previous research, are for 3D visualization of the roadside standing building scene captured by a side-looking video camera as a continuous set of vertical-textured planar faces. These vertical faces were concatenated to create an approximate model on which the images could be back-projected as textures. We proposed a retrieval technique of selecting vertical-textured planar faces from archived 3D image mosaics to fast represent a 3D world, where this selection is based on proximity to the user location in the virtual world. The chosen vertical-textured planar faces are recomposed as a continuous set formed stream and are then rendered in 3D virtual space.

Recently, motion capture is widely used in the human animation. But it is still difficult to create new-style animations from the existing motion capture data. In this paper, we propose a novel technique to create the style animations from an existing motion sequence. Firstly, we use linear time-invariant system (LTI) to derive an explicit mapping between the high-dimensional motion capture data and the low-dimensional state variables. Secondly, new style animations are created within the state space. Only a few important keyframes need be modified through adjusting the low-dimensional style variables. The remaining frames of the original motion can be generated automatically. Finally, we design an effective algorithm to calculate the model parameters. Experimental results show that the generated style animations are natural and smooth.

Employing computer techniques to visualize the color changing of Dunhuang Frescoes is of great significance since the audience can re-experience this process and the protectors can discuss the protection measures according to the simulated video clips and thus help them make their decision more scientifically. But previously, because colorimetric and chemical modeling was not introduced to simulate the color changing of frescoes, there was a lack of objectivity and sense of reality. In this paper, however, two kinds of color changing, namely, color changing due to chemical changing of the pigment as well as dust accumulation, are modeled. Since this modeling describes the natural behavior of the color changing process, the simulation result is quite striking.

The mechanical behaviour of the human skin shows a non-linear stress-strain relationship. Based on the behaviour of human dermis proposed by Hendriks, we develop a system to draw a specific skin incision shape which consists of four spine curves and two lines. Users can firstly define the two incision points on the skin, our system can superimpose the shape on the surface and show the skin incision cutaway. OBBTree analysis for the skin surface is used to put the incision shape. Followed by shape shown on the surface and finally a cutaway is generated to show the cut. Jittering problem generated by the above steps is also resolved.

We present a framework for generating paper-cutting effects on mesh models. Our system involves the construction of model defined paper-cutting patterns, extraction of the medial axis (MA) from the mesh models and position location for patterns on the mesh surface based on MA. Patch sets for placing patterns on the mesh surface are determined by the region growing algorithm. After patch sets parameterization, patterns are drawn on the parameterized domain and mapped onto the mesh surface by inverse sampling.

In this paper, we present a new dynamic texture technique by using our new mixed auto-regressive moving average exogenous (MARMAX) model in the wavelet domain. The technique captures much more detailed local and global dynamic texture properties than the previous works and it can synthesize long dynamic texture videos with temporal coherency. The whole input video is firstly transformed into the wavelet domain in different levels, then we use MARMAX model to capture the dynamic characters of the original video, finally we transform the video back by inverse discrete wavelet transform (IDWT). A desired long dynamic texture video can be synthesized easily from a short input one. The experimental results demonstrate that our approach can produce visually promising dynamic texture sequences.

Regarding the modeling aspects of computer animation, the skeletal animation is a newfangled modeling technology to assist in single-mesh model animation. There are two aspects of advancement proposed in this paper. First, we construct a novel vertical data structure (Modified Body-Tree) to describe the skeletal models in animation. Second, instead of normal methods such as linear interpolation and higher-order spline interpolation to generate in-between frame, we proposes an adaptive and high-performance algorithm based on 4-Point subdivision scheme to generate in-between frame in the skeletal animation. By this advanced algorithm, a smoother skeletal animation with blander vision effect is achieved.

Virtual Five Animals Exercise is an interactive game which is an important part of Virtual Olympic Museum for the Games of the XXIX Olympiad. Users can learn and imitate Five Animals Exercise without any interactive device. To do these, we make use of motion capture data to drive the virtual human motion, and use template matching method to recognize users’ motion. We improve the motion editing and motion retargetting methods in our system. Many people enjoy our game.

In this paper, we formulate the image matting as one of solving energy minimization problems. Our approach has the following advantages. First, the matte estimation is modeled using an energy function as a MRF optimization problem. Second, the energy function combines the gradient of the matte, the gradient of the color and statistical sampling together to achieve global optimization. Third, the matte is directly reconstructed by solving energy equations. Experimental results show that our method is efficient to extract high quality mattes for foregrounds with complex natural images.

This paper describes the design of a distributed architecture to support an interactive, interoperable, and collaborative forest fire fighting simulation using High Level Architecture (HLA). Based on the Run-time Infrastructure (RTI) services which are specified in HLA and C++ application programming interface (API) of the RTI, The distributed virtual fire fighting environment provides a practical foundation to enhance interactivity, interoperability for distributed simulation.. Users can build new federation application by the way of the goal system reorganization. The key techniques, such as FOM/SOM design, system structure, running mechanism of simulation system are discussed.