Claudio Esperança

Use of the SAND spatial browser for digital government applications

Numerous federal agencies produce official statistics made accessible to ordinary citizens for searching and data retrieval. This is frequently done via the Internet through a Web browser interface. If this data is presented in textual format, it can often be searched and retrieved by such attributes as topic, responsible agency, keywords, or press release. However, if the data is of spatial nature, for example, in the form of a map, then using text-based queries is often too cumbersome for the intended audience. We describe the use of the SAND Spatial Browser to provide more power to users of these databases by enabling them to define and explore the specific spatial region of interest graphically. The SAND Spatial Browser allows users to form either purely spatial or mixed spatial/nonspatial queries intuitively, which can present information to users that might have been missed if only a textual interface was available.

Volume and Isosurface Rendering with GPU-Accelerated Cell Projection*

We present an efficient Graphics Processing Unit GPU‐based implementation of the Projected Tetrahedra (PT) algorithm. By reducing most of the CPU–GPU data transfer, the algorithm achieves interactive frame rates (up to 2.0 M Tets/s) on current graphics hardware. Since no topology information is stored, it requires substantially less memory than recent interactive ray casting approaches. The method uses a two‐pass GPU approach with two fragment shaders. This work includes extended volume inspection capabilities by supporting interactive transfer function editing and isosurface highlighting using a Phong illumination model.

Orthogonal Polygons as Bounding Structures in Filter-Refine Query Processing Strategies

The use of bounding structures in the form of orthogonal polygons (also known as rectilinear polygons) with a varying number of vertices in contrast with a minimum bounding rectangle (an orthogonal polygon with just 4 vertices in two dimensions) as an object approximation method is presented. Orthogonal polygons can be used to improve the performance of the refine step in the filter-refine query processing strategy employed in spatial databases. The orthogonal polygons are represented using the vertex representation implemented as a vertex list. The advantage of the vertex representation implemented as a vertex list is that it can be used to represent orthogonal polygons in arbitrary dimensions using just their vertices. This is in contrast to conventional methods such as the chain code which only work in two dimensions and cannot be extended to deal with higher dimensional data. Algorithms are given for varying the number of vertices used to represent the objects. It is shown that the use of non-trivial orthogonal polygons (i.e., with more than four vertices) is of benefit when a spatial index is used in the filter step for processing spatial queries such as point-in-object and windowing. If no spatial index is used, then all objects must be examined. In this case, many of the objects are small thereby not benefiting from the variation in the number of vertices that they have as the simple bounding box is adequate.

GPU-Based Cell Projection for Interactive Volume Rendering

We present a practical approach for implementing the projected tetrahedra (PT) algorithm for interactive volume rendering of unstructured data using programmable graphics cards. Unlike similar works reported earlier, our method employs two fragment shaders, one for computing the tetrahedra projections and another for rendering the elements. We achieve interactive rates by storing the model in texture memory and avoiding redundant projections of implementations using vertex shaders. Our algorithm is capable of rendering over 2.0 M Tet/s on current graphics hardware, making it competitive with recent ray-casting approaches, while occupying a substantially smaller memory footprint

A Collision Detection and Response Scheme for Simplified Physically Based Animation

In this paper we describe a system for physical animation of rigid and deformable objects. These are represented as groups of particles linked by linear constraints, while a Verlet integrator is used for motion computation. Unlike traditional approaches, we accomplish physical simulation without explicitly computing orientation matrices, torques or inertia tensors. The main contribution of our work is related to the way collisions are handled by the system, which employs different approaches for deformable and rigid bodies. In particular, we show how collision detection using the GJK algorithm [9] and bounding sphere hierarchies can be combined with the projection based collision response technique described by Jakobsen [14].

A differential code for shape representation in image database applications

A new method termed the vertex representation is presented for approximating the shapes of objects of arbitrary dimensionality d (e.g., 2D, 3D, etc.) with orthogonal (d-1)-dimensional faces using a variable number of vertices. The vertex representation can be viewed as a generalization of boundary codes (i.e., chain codes) to higher dimensions. Techniques are described for using the vertex representation to efficiently perform many operations commonly performed on rasters. The utility of these techniques in image database applications is discussed.

A progressive vector map browser for the web

With the increasing popularity of web-based map browsers, remotely obtaining a high quality depiction of cartographic information has become commonplace. Most web mapping systems, however, rely on high-capacity servers transmitting pre-rendered tiled maps in raster format. That approach is capable of producing good quality renderings on the client side while using limited bandwidth and exploiting the browser’s image cache. These goals are harder to achieve for maps in vector format. In this work, we present an alternative client-server architecture capable of progressively transmitting vector maps in levels-of-detail (LOD) by using techniques such as polygonal line simplification, spatial data structures and, most importantly, a customized memory management algorithm. A multiplatform implementation of this system is described, where the client application is written entirely in JavaScript and processed within the web browser, avoiding the need of external applications or plug-ins. Results of experiments aimed at gauging both the performance and the display quality obtained with the system are presented and explained. Extensions to the system are also discussed, including issues such as level-of-detail versus visual importance tradeoffs and the handling of closed polygonal lines.

Enhancing the Bayesian network approach to face detection

This work presents some improvements to the Bayesian network approach to face detection. Feature selection is directed towards avoiding the problems caused by considering feature fragments as complete features. More selective criteria for grouping are applied, thus reducing the number of candidate groups that must be considered. Statistical measures dependent on the feature size are used to increase the detection reliability, especially in images of groups of people.

Loop snakes: the generalized model

The topologically adaptable snake model, or simply T-snakes, is a useful tool for automatically identifying multiple segments in an image. According to A. Oliveira et al. (2004), a novel approach for controlling the topology of a T-snake was introduced. That approach focuses on the loops formed by the projected curve which is obtained at every stage of the snake evolution. The idea is to make that curve the image of a piecewise linear mapping of an adequate class. Then, with the help of an additional structure, the loop-tree, it is possible to decide in 0(1) time whether the region delimited by each loop has already been explored by the snake. In the original proposal of the loop snakes model, the snake evolution is limited to contraction and there is only one T-snake that contracts and splits during evolution. In this paper we generalize the original model by allowing the contraction as well as the expansion of several T-snakes.

Loop snakes: snakes with enhanced topology control

Topologically adaptable snakes, or simply T-snakes, are a standard tool for automatically identifying multiple segments in an image. This work introduces a novel approach for controlling the topology of a T-snake. It focuses on the loops formed by the so-called projected curve which is obtained at every stage of the snake evolution. The idea is to make that curve the image of a piecewise linear mapping of an adequate class. Then, with the help of an additional structure - the loop-tree - it is possible to decide in O(1) time whether the region enclosed by each loop has already been explored by the snake. This makes it possible to construct an enhanced algorithm for evolving T-snakes whose performance is assessed by means of statistics and examples.