Catálogo de publicaciones - libros

The extraction of a parametric global motion from a motion field is a task with several applications in video processing. We present two probabilistic formulations of the problem and carry out optimization using the RAST algorithm, a geometric matching method novel to motion estimation in video. RAST uses an exhaustive and adaptive search of transformation space and thus gives – in contrast to local sampling optimization techniques used in the past – a globally optimal solution. Among other applications, our framework can thus be used as a source of ground truth for benchmarking motion estimation algorithms.

Our main contributions are: first, the novel combination of a state-of-the-art MAP criterion for dominant motion estimation with a search procedure that guarantees global optimality. Second, experimental results that illustrate the superior performance of our approach on synthetic flow fields as well as real-world video streams. Third, a significant speedup of the search achieved by extending the model with an additional smoothness prior.

Variational methods are among the most successful approaches to calculate the optical flow between two image frames. A particularly appealing formulation is based on total variation (TV) regularization and the robust norm in the data fidelity term. This formulation can preserve discontinuities in the flow field and offers an increased robustness against illumination changes, occlusions and noise. In this work we present a novel approach to solve the TV- formulation. Our method results in a very efficient numerical scheme, which is based on a dual formulation of the TV energy and employs an efficient point-wise thresholding step. Additionally, our approach can be accelerated by modern graphics processing units. We demonstrate the real-time performance (30 fps) of our approach for video inputs at a resolution of 320×240 pixels.

Magnetic resonance spectral images provide information on metabolic processes and can thus be used for in vivo tumor diagnosis. However, each single spectrum has to be checked manually for tumorous changes by an expert, which is only possible for very few spectra in clinical routine. We propose a semi-supervised procedure which requires only very few labeled spectra as input and can hence adapt to patient and acquisition specific variations. The method employs a discriminative random field with highly flexible single-side and parameter-free pair potentials to model spatial correlation of spectra. Classification is performed according to the label set that minimizes the energy of this random field. An iterative procedure alternates a parameter update of the random field using a kernel density estimation with a classification by means of the GraphCut algorithm. The method is compared to a single spectrum approach on simulated and clinical data.

The ability of a segmentation algorithm to uncover an interesting partition of an image critically depends on its capability to utilize and combine all available, relevant information. This paper investigates a method to automatically weigh different data sources, such that a meaningful segmentation is uncovered. Different sources of information naturally arise in image segmentation, e.g. as intensity measurements, local texture information or edge maps. The data fusion is controlled by a regularization mechanism, favoring sparse solutions. Regularization parameters as well as the clustering complexity are determined by the concept of cluster stability yielding maximally reproducible segmentations. Experiments on the Berkeley segmentation database show that our segmentation approach outperforms competing segmentation algorithms and performs comparably to supervised boundary detectors.

This paper presents a bottom-up approach for fast segmentation of natural images. This approach has two main stages: firstly, it detects the homogeneous regions of the input image using a colour-based distance and then, it merges these regions using a more complex distance. Basically, this distance complements a contrast measure defined between regions with internal region descriptors and with attributes of the shared boundary. These two stages are performed over the same hierarchical framework: the Bounded Irregular Pyramid (BIP). The performance of the proposed algorithm has been quantitatively evaluated with respect to ground-truth segmentation data.

Image segmentation based on pairwise pixel similarities has been a very active field of research in recent years. The drawbacks common to these segmentation methods are the enormous space and processor requirements. The contribution of this paper is a general purpose two-stage preprocessing method that substantially reduces the involved costs. Initially, an oversegmentation into small coherent image patches - or superpixels - is obtained through an iterative process guided by pixel similarities. A suitable pairwise superpixel similarity measure is then defined which may be plugged into an arbitrary segmentation method based on pairwise pixel similarities. To illustrate our ideas we integrated the algorithm into a spectral graph-partitioning method using the Normalized Cut criterion. Our experiments show that the time and memory requirements are reduced drastically (> 99%), while segmentations of adequate quality are obtained.

Autonomous collision avoidance in vehicles requires an accurate separation of obstacles from the background, particularly near the focus of expansion. In this paper, we present a technique for fast segmentation of stationary obstacles from video recorded by a single camera that is installed in a moving vehicle. The input image is divided into three motion segments consisting of the ground plane, the background, and the obstacle. This constrained scenario allows for good initial estimates of the motion models, which are iteratively refined during segmentation. The horizon is known due to the camera setup. The remaining binary partitioning problem is solved by a graph cut on the motion-compensated difference images.

Obstacle segmentation in realistic scenes with a monocular camera setup has not been feasible up to now. Our experimental evaluation shows that the proposed approach leads to fast and accurate obstacle segmentation and distance estimation without prior knowledge about the size, shape or base point of obstacles.

Research on EEG based brain-computer-interfaces (BCIs) aims at steering devices by thought. Even for simple applications, BCIs require an extremely effective data processing to work properly because of the low signal-to-noise-ratio (SNR) of EEG signals. Spatial filtering is one successful preprocessing method, which extracts EEG components carrying the most relevant information. Unlike spatial filtering with Common Spatial Patterns (CSP), Adaptive Spatial Filtering (ASF) can be adapted to freely selectable regions of interest (ROI) and with this, artifacts can be actively suppressed. In this context, we compare the performance of ASF with ROIs selected using anatomical a-priori information and ASF with numerically optimized ROIs. Therefore, we introduce a method for data driven spatial filter adaptation and apply the achieved filters for classification of EEG data recorded during imaginary movements of the left and right hand of four subjects. The results show, that in the case of artifact-free datasets, ASFs with numerically optimized ROIs achieve classification rates of up to 97.7 % while ASFs with ROIs defined by anatomical heuristic stay at 93.7 % for the same data. Otherwise, with noisy datasets, the former brake down (66.7 %) while the latter meet 95.7 %.

We describe a technique to detect and localize features on checkerboard calibration charts with high accuracy. Our approach is based on a model representing the sought features by a multiplicative combination of two edge functions, which, to allow for perspective distortions, can be arbitrarily oriented.

First, candidate regions are identified by an eigenvalue analysis of the structure tensor. Within these regions, the sought checkerboard features are then detected by matched filtering. To efficiently account for the double-oriented nature of the sought features, we develop an extended version of steerable filters, viz., multi-steerable filters. The design of our filters is carried out by a Fourier series approximation. Multi-steerable filtering provides both the unknown orientations and the positions of the checkerboard features, the latter with pixel accuracy. In the last step, the feature positions are refined to subpixel accuracy by fitting a paraboloid. Rigorous comparisons show that our approach outperforms existing feature localization algorithms by a factor of about three.

Thinning on binary images is an iterative layer by layer erosion until only the “skeletons” of the objects are left. This paper presents an efficient parallel 3D surface–thinning algorithm. A three–subiteration strategy is proposed: the thinning operation is changed from iteration to iteration with a period of three according to the three deletion directions.