Werner von Seelen

AN IMAGE PROCESSING SYSTEM FOR DRIVER ASSISTANCE

In this paper, the authors describe a system designed to extract information from an image acquired from an onboard CCD camera. The purpose of the system is to detect, track and classify objects. An approach that involves integration and fusion in the sequential and parallel phases of sensor and information processing is described. The authors note that the primary advantage of this approach is the integrative coupling of different algorithms providing partly redundant information.

Intensity and Edge-Based Symmetry Detection Applied to Car-Following

We present two methods for detecting symmetry in images, one based directly on the intensity values and another one based on a discrete representation of local orientation. A symmetry finder has been developed which uses the intensity-based method to search an image for compact regions which display some degree of mirror symmetry due to intensity similarities across a straight axis. In a different approach, we look at symmetry as a bilateral relationship between local orientations. A symmetryenhancing edge detector is presented which indicates edges dependent on the orientations at two different image positions. SEED, as we call it, is a detector element implemented by a feedforward network that holds the symmetry conditions. We use SEED to find the contours of symmetric objects of which we know the axis of symmetry from the intensity-based symmetry finder. The methods presented have been applied to the problem of visually guided car-following. Real-time experiments with a system for automatic headway control on motorways have been successful.

Computer vision for driver assistance systems

Systems for automated image analysis are useful for a variety of tasks and their importance is still increasing due to technological advances and an increase of social acceptance. Especially in the field of driver assistance systems the progress in science has reached a level of high performance. Fully or partly autonomously guided vehicles, particularly for road-based traffic, pose high demands on the development of reliable algorithms due to the conditions imposed by natural environments. At the Institut fur Neuroinformatik, methods for analyzing driving relevant scenes by computer vision are developed in cooperation with several partners from the automobile industry. We introduce a system which extracts the important information from an image taken by a CCD camera installed at the rear view mirror in a car. The approach consists of a sequential and a parallel sensor and information processing. Three main tasks namely the initial segmentation (object detection), the object tracking and the object classification are realized by integration in the sequential branch and by fusion in the parallel branch. The main gain of this approach is given by the integrative coupling of different algorithms providing partly redundant information.

Optical Imaging of Cat Auditory Cortex Cochleotopic Selectivity Evoked by Acute Electrical Stimulation of a Multi‐channel Cochlear Implant

We measured reflectance changes by means of optical imaging of intrinsic signals to study the effects of acute electrical cochlear stimulation on the topography of the cat auditory cortex. After single‐pulse electrical stimulation at selected sites of a multichannel implant device, we found topographically restricted response areas representing mainly the high‐frequency range in Al. Systematic variation of the stimulation pairs and thus of the cochlear frequency sites revealed a systematic and corresponding shift of the response areas that matched the underlying frequency organization. Intensity functions were usually very steep. Increasingly higher stimulation currents evoked increasingly larger response areas, resulting in decreasing spatial, i.e. cochleotopic, selectivity; however, we observed only slight positional shifts of the focal zones of activity. Electrophysiological recordings of local field potential maps in the same individual animals revealed close correspondence of the locations of the cortical response areas. The results suggest that the method of optical imaging can be used to map response areas evoked by electrical cochlear stimulation, thereby maintaining a profound cochleotopic selectivity. Further experiments in chronically stimulated animals will shed more light on the degree of functional and reorganizational capacities of the primary auditory cortex and could be beneficial for our understanding of the treatment of profound deafness.

Neural mapping and space-variant image processing

Abstract The mapping of sensory surfaces onto cortical areas as well as maps in inter-areal projections are classified and discussed in terms of space-variant image processing. The neurobiological phenomena discussed in this context include retinotopic maps, columnar organization, interactions between a retinotopic map and a uniform operation in a cortical area, and patchy or segregated projections of two or more areas to a common target area. We present a mathematical framework for describing the interaction of neural mappings and local image processing operations which allows functional interpretations. In an example from visual navigation, we show that neural maps are powerful tools for the parallel processing of visual information. Since mappings of the various types result in a spatial encoding of arbitrary stimulus information, image processing operations can be applied to information from other modalities or for higher-level problems as well. For this aspect of neural mapping, we adopt the term parametric mapping.

Optical imaging of rat somatosensory cortex reveals representational overlap as topographic principle

We measured reflectance changes by means of optical imaging of intrinsic signals to study the topography of the paw representations in rat somatosensory cortex. Following circumscribed tactile stimulation of single digits or pads, we found large and partially overlapping areas of reflectance changes (ΔR). The diamaters of their focal zones defined at 75% maximal ΔR were in the range of 150 µm and preserved all details of the underlying maps. Zones of overlap were in the range 15–25% measured at half-maximal ΔR. In contrast, we found sharp boundaries with no overlap between the fore- and hindpaw representations. The data suggest that large and overlapping cortical maps constitute a normal type of neural representation supporting the idea of a distributed neural processing scheme.

Entropie als Maß des lokalen Informationsgehalts in Bildern zur Realisierung einer Aufmerksamkeitssteuerung

Basierend auf der Informationstheorie, die C. Shannon [Sha48] einfuhrte, wird der lokale Informationsgehalt in Bildern geschatzt. Hierbei wird die Entropie als Mas der zu erwartenden Information eines Bildausschnitts herangezogen. Dieses lokale Bildentropiemas realisiert eine Aufmerksamkeitssteuerung, die ein Teilmodul der Anwendung „Autonomes Fuhren von Fahrzeugen” bildet.

Fourier optical approach to the extraction of morphological parameters from the diffraction pattern of biological cells

A new procedure for fast quantitative extraction of cell parameters from diffraction patterns was deduced from model calculations and applied to cervical gynecological material. Based on radial scans of the diffraction pattern, the technique permits simultaneous determination of the nuclear and cytoplasmic diameters by Fourier analysis of the radial scanning signals after compensation for the intensity falloff by an amplification proportional to the third power of the radial position in the diffraction plane. Illustrative examples of measurements on exfoliated cells of different types are presented.

Complex adaptation and system structure.

The structural organization of biological systems is one of natures most fascinating aspects, but its origin and functional role is not yet fully understood. For instance, basic adaptational mechanisms like genetic mutation and Hebbian adaptation seem to be generic and invariant across many species and are, on their own, fairly well investigated and understood. However, it is the organisms structure - the representations these mechanisms act upon - that bears the complex functional effects of these mechanisms. While typical technical approaches to system design require detailed problem models and suffer from the need to explicitly take care of all possible cases, the organization of biological systems seems to induce inherent adaptability, flexibility and robustness. In this discussion paper we address the concept of structured variability, particularly the role of system structure as implementing a certain representation on which basic variational mechanisms act on. The functional adaptability (or search distribution) depends crucially on this representation.

Evaluating flexible fuzzy controllers via evolution strategies

This paper applies evolution strategies to evaluate a class of flexible fuzzy controllers. Various components in fuzzy controllers, including generalized T-norms, soft T-norm, BADD defuzzification and scaling factors are investigated. Simulations are carried out on linear and nonlinear systems and some interesting conclusions are obtained. These conclusions are further explained with an application example of vehicle control. It is hoped that they will be helpful for design of fuzzy controllers with better control performance.