Learning compatibility functions for feature binding and perceptual grouping

Authors:
Sebastian Weng;Jochen J. Steil
Affiliations:
Bielefeld University, Neuroinformatics Department, Bielefeld, Germany;Bielefeld University, Neuroinformatics Department, Bielefeld, Germany
Venue:
ICANN/ICONIP'03 Proceedings of the 2003 joint international conference on Artificial neural networks and neural information processing
Year:
2003

Citing 6
Cited 2

Global competition and local cooperation in a network of neural oscillators

Physica D
Efficient Vector Quantization Using the WTA-Rule with Activity Equalization

Neural Processing Letters
Maximum Contrast Classifiers

ICANN '02 Proceedings of the International Conference on Artificial Neural Networks
Data Driven Generation of Interactions for Feature Binding and Relaxation Labeling

ICANN '02 Proceedings of the International Conference on Artificial Neural Networks
A Competitive-Layer Model for Feature Binding and Sensory Segmentation

Neural Computation
A neural classifier enabling high-throughput topological analysis of lymphocytes in tissue sections

IEEE Transactions on Information Technology in Biomedicine

Integrating feature maps and competitive layer architectures for motion segmentation

Neurocomputing
Perspectives of self-adapted self-organizing clustering in organic computing

BioADIT'06 Proceedings of the Second international conference on Biologically Inspired Approaches to Advanced Information Technology

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present and compare data driven learning methods to generate compatibility functions for feature binding and perceptual grouping. As dynamic binding mechanism we use the competitive layer model (CLM), a recurrent neural network with linear threshold neurons. We introduce two new and efficient learning schemes and also show how more traditional standard approaches as MLP or SVM can be employed as well. To compare their performance, we define a measure of grouping quality with respect to the available training data and apply all methods to a set of real world fluorescence cell images.