Catálogo de publicaciones - libros

The subjective contour generation in the human brain depends on the interaction of local and comprehensive processes realize this mechanism. The purpose of this paper is to propose a model that outputs subjective contours from line figures. This model, for a local process, detects endpoints and generates potential fields that represent probability of an occluding contour’s existence. Each field is anisotropic and spreads perpendicularly to the line figure. Then, for a comprehensive process, the directions of potential fields are corrected to the degree their intersections are maximized in the image. Finally, it fixes subjective contours by tracking potential ridgelines and outputs a result image. The generated subjective contour is smoothly curved and the shape is appropriate compare to what we perceive.

A method is presented for constructing a set of triangles that closely approximates the surface of a constructive solid geometry model. The method subdivides an initial triangulation of the model’s primitives into triangles that can be classified accurately as either on or off of the surface of the whole model, and then recombines these small triangles into larger ones that are still either entirely on or entirely off the surface. Subdivision and recombination can be done in a preprocessing step, allowing later rendering of the triangles on the surface (, the triangles visible from outside the model) to proceed at interactive rates. Performance measurements confirm that this method achieves interactive rendering speeds. This approach has been used with good results in an interactive scientific visualization program.

Visualization of glyphs has a long history in medical imaging but gains much more power when the glyphs are properly placed to fill the screen. Glyph packing is often performed via an iterative approach to improve the location of glyphs. We present an alternative implementation of glyph packing based on a Delaunay triangulation to speed up the clustering process and reduce costs for neighborhood searches. Our approach does not require a re–computation of acceleration structures when a plane is moved through a volume, which can be done interactively. We provide two methods for initial placement of glyphs to improve the convergence of our algorithm for glyphs larger and glyphs smaller than the data set’s voxel size. The main contribution of this paper is a novel approach to glyph packing that supports simpler parameterization and can be used easily for highly efficient interactive data exploration, in contrast to previous methods.

This paper presents a method to reduce time complexity of the computation of higher–order tensor lines. The method can be applied to higher–order tensors and the spherical harmonics representation, both widely used in medical imaging. It is based on a gradient descend technique and integrates well into fiber tracking algorithms. Furthermore, the method improves the angular resolution in contrast to discrete sampling methods which is especially important to tractography, since there, small errors accumulate fast and make the result unusable. Our implementation does not interpolate derived directions but works directly on the interpolated tensor information. The specific contribution of this paper is a fast algorithm for tracking lines tensor fields of arbitrary order that increases angular resolution compared to previous approaches.

We present a method to speed up the computation of a high-quality data-dependent triangulation approximating an image using simulated annealing by probability distributions guided by local approximation error and its variance. The triangulation encodes the image, yielding compression rates comparable to or even superior to JPEG and JPEG2000 compression.

The specific contributions of our paper are a speed-up of the simulated annealing optimization and a comparison of our approach to other image approximation and compression methods. Furthermore, we propose an adaptive vertex insertion/removal strategy and termination criteria for the simulated annealing to achieve specified approximation error bounds.

We present a novel technique for optical data transfer between public displays and mobile devices based on unsynchronized 4D barcodes. We assume that no direct (electromagnetic or other) connection between the devices can exist. Time-multiplexed, 2D color barcodes are displayed on screens and recorded with camera equipped mobile phones. This allows to transmit information optically between both devices. Our approach maximizes the data throughput and the robustness of the barcode recognition, while no immediate synchronization exists. Although the transfer rate is much smaller than it can be achieved with electromagnetic techniques (e.g., Bluetooth or WiFi), we envision to apply such a technique wherever no direct connection is available. 4D barcodes can, for instance, be integrated into public web-pages, movie sequences, advertisement presentations or information displays, and they encode and transmit more information than possible with single 2D or 3D barcodes.

A major challenge in deploying service robots into the real world is to design a framework that provides effective, long-term interactions with people. This includes interacting with people in a natural way, dealing with multiple users, and being continually aware of the surroundings. This paper proposes a robot control architecture that addresses these issues. First, it enables the representation of complex, sequential, and hierarchical robot tasks, in a behavior-based framework. Second, it provides a robot with the flexibility to deal with multiple requests and interruptions, over extended periods. Third, it uses a visual awareness mechanism to recognize users and to identify their need for robot interaction. We demonstrate our approach on a Pioneer 3DX mobile robot, performing service tasks in a real-world environment.

The aim of this study is to classify structural cartographic objects in high-resolution satellite images. The target classes have an important intra-class variability because the class definitions belong to high-level concepts. Structural attributes seem to be the most plausible cues for the classification task. We propose an Adaboost learning method using edge-based features as weak learners. Multi-scale sub-pixel edges are converted to geometrical primitives as potential evidences of the target object. A feature vector is calculated from the primitives and their perceptual groupings, by the accumulation of combinations of their geometrical and spatial attributes. A classifier is constructed using the feature vector. The main contribution of this paper is the usage of structural shape attributes in a statistical learning method framework.

We tested our method on CNES dataset prepared for the ROBIN Competition and we obtained promising results.

Dust emissions due to low-level flight of a helicopter are studied as part of the Integrated Desert Terrain Forecasting for Military Operations project. Atmospheric concentrations of were measured at different heights downwind of the helicopter flight path. Digital video images captured the entrainment and dispersion of the dust plume formed by the wake of the helicopter during each pass down the flight course. The video data are analyzed to relate the dust plume strength to degradation of local visibility. A strong relationship between color changes/standard deviations and plume strength is found. This relationship is used to develop an algorithm that can determine local visibility degradation due to local concentrations around a helicopter. This algorithm can be combined with concentration output data from an atmospheric dispersion model to simulate visibility in a helicopter simulator.

The ultimate print quality evaluation is always based on end-users’ “quality experience”, and therefore, the main challenge in automatic evaluation is to model the visual path and cognition process from physical properties to the experience. The present efforts to automate print quality evaluation have been concentrated on automating the current manually-performed assesments, which reduces the laborious work, but does not provide any novel information. In this work, a new approach for automating the evaluation is proposed and the approach is utilised by defining new computational level features which are able to explain visual quality evaluations performed by human experts.