Gabriele Peters

Robust detection of significant points in multiframe images

Significant point (SP) detection is an important pre-processing step in image registration, data fusion, object recognition and in many other tasks. This paper deals with multiframe SP detection, i.e. detection in two or more images of the same scene which are supposed to be blurred, noisy, rotated and shifted with respect to each other. We present a new method invariant under rotation that can handle diAerently blurred images. Thanks to this, the point sets extracted from diAerent frames have relatively high number of common elements. This property is highly desirable for further multiframe processing. The performance of the method is demonstrated experimentally on satellite images. ” 1999 Elsevier Science B.V. All rights reserved.

Aesthetic Primitives of Images for Visualization

Images play an important role in visualization. As users are more willing to adopt a product if it evokes pleasurable feelings the aesthetic appeal of interfaces becomes more important. Thus, there is a growing need to generate also images which appear aesthetically to the user. Starting with the modularities of the human visual system, we derive six dimensions of visual aesthetics. For each dimension we explore, inspired by principles of the visual arts and insights of cognitive neuroscience, which pecularities of the dimensions are particularly adequate for an aesthetic impression. Accompanied by a fair number of image examples, these considerations result in an easy to understand guideline for computer scientists and interface designers how to deal with images in terms of aesthetics.

View Reconstruction by Linear Combination of Sample Views

Ullman and Basri [1] have shown theoretically, that a three-dimensional object can be represented by a linear combination of two-dimensional images of the object. But they have applied their calculations to artificially created images only, like line drawings of cars. The application to images of real objects turns out to be difficult, because a crucial point of their algorithm is the knowlegde of correspondences in the sample views. In this article we describe a biologically inspired system which automatically provides correspondences between views of a three-dimensional object. This enables us to apply Ullman and Basri’s linear combination approach to images of arbitrary, real objects. We give detailed formula of our linear combinations and examples for reconstructed object views.

Information Theory and the Brain: Principles of Cortical Processing Applied to and Motivated by Artificial Object Recognition

In this paper we discuss the biological plausibility of the object recognition system described in detail in (Kruger, Peters and v.d. Malsburg, 1996). We claim that this system realizes the following principles of cortical processing: hierarchical processing, sparse coding, and ordered arrangement of features. Furthermore, our feature selection is motivated by response properties of neurons in striate cortex and by Biederman’s theory of object representation on higher stages of visual processing (Biederman, 1987). Inspired by the current discussion about aspects of cortical processing, we hope to derive more efficient algorithms. By discussing the functional meaning of these aspects in our object recognition system, we hope to attain a deeper understanding of their meaning for brain processing.

A New View on Grid Cells Beyond the Cognitive Map Hypothesis

Grid cells in the entorhinal cortex are generally considered to be a central part of a path integration system supporting the construction of a cognitive map of the environment in the brain. Guided by this hypothesis existing computational models of grid cells provide a wide range of possible mechanisms to explain grid cell activity in this specific context. Here we present a complementary grid cell model that treats the observed grid cell behavior as an instance of a more abstract, general principle by which neurons in the higher-order parts of the cortex process information.

A Vision System for Interactive Object Learning

We propose an architectural model for a responsive vision system based on techniques of reinforcement learning. It is capable of acquiring object representations based on the intended application. The system can be interpreted as an intelligent scanner that interacts with its environment in a perception-action cycle, choosing the camera parameters for the next view of an object depending on the information it has perceived so far. The main contribution of this paper consists in the presentation of this general architecture which can be used for a variety of applications in computer vision and computer graphics. In addition, the funcionality of the system is demonstrated with the example of learning a sparse, view-based object representation that allows for the reconstruction of non-acquired views. First results suggest the usability of the proposed system.

An alternative approach to the revision of ordinal conditional functions in the context of multi-valued logic

We discuss the use of Ordinal Conditional Functions (OCF) in the context of Reinforcement Learning while introducing a new revision operator for conditional information. The proposed method is compared to the state-of-the-art method in a small Reinforcement Learning application with added futile information, where generalization proves to be advantageous.

ORASSYLL: Object Recognition with Autonomously Learned and Sparse Symbolic Representations Based on Metrically Organized Local Line Detectors

Abstract We introduce an object recognition and localization system in which objects are represented as a sparse and spatially organized set of local (bent) line segments. The line segments correspond to binarized Gabor wavelets or banana wavelets, which are bent and stretched Gabor wavelets. These features can be metrically organized; the metric enables an efficient learning of object representations. It is essential for learning that only corresponding local areas are compared with each other; i.e., the correspondence problem has to be solved. We achieve correpondence (and in this way autonomous learning) by utilizing motor-controlled feedback, i.e., by interaction of arm movement and camera tracking. The learned representations are used for fast and efficient localization and discrimination of objects in complex scenes.

Analysis of high-dimensional data using local input space histograms

Abstract The idea of local input space histograms was recently introduced as a means to augment prototype-based vector quantization methods in order to gather more information about the structure of the respective input space. Here we investigate the utility of this new idea for analysing and clustering high-dimensional data. Our results demonstrate that the additional information gained about the input space structure can be used to enable and improve visualization and hierarchical clustering. Furthermore, we show that contrary to common view the Minkowski distance with p > 1 can be a meaningful distance measure for high-dimensional data.

Ranking Functions in Large State Spaces

Large state spaces pose a serious problem in many learning applications. This paper discusses a number of issues that arise when ranking functions are applied to such a domain. Since these functions, in their original introduction, need to store every possible world model, it seems obvious that they are applicable to small toy problems only. To disprove this we address a number of these issues and furthermore describe an application that indeed has a large state space. It is shown that an agent is enabled to learn in this environment by representing its belief state with a ranking function. This is achieved by introducing a new entailment operator that accounts for similarities in the state description.