Distortion Invariant Object Recognition in the Dynamic Link Architecture

@article{Lades1993DistortionIO,
  title={Distortion Invariant Object Recognition in the Dynamic Link Architecture},
  author={Martin Lades and Jan C. Vorbr{\"u}ggen and Joachim M. Buhmann and J{\"o}rg Lange and Christoph von der Malsburg and Rolf P. W{\"u}rtz and Wolfgang Konen},
  journal={IEEE Trans. Computers},
  year={1993},
  volume={42},
  pages={300-311}
}
An object recognition system based on the dynamic link architecture, an extension to classical artificial neural networks (ANNs), is presented. The dynamic link architecture exploits correlations in the fine-scale temporal structure of cellular signals to group neurons dynamically into higher-order entities. These entities represent a rich structure and can code for high-level objects. To demonstrate the capabilities of the dynamic link architecture, a program was implemented that can recognize… 

Figures and Tables from this paper

Face recognition by dynamic link matching

TLDR
A neural system for the recognition of objects from realistic images, together with results of tests of face recognition from a large gallery, based on Dynamic Link Matching, which requires very little genetic or learned structure.

Combining Feature- and Correspondence-Based Methods for Visual Object Recognition

TLDR
An object recognition system built on a combination of feature- and correspondence-based pattern recognizers that achieved high recognition rates on identity and pose and can also cope with varying background, multiple objects, and partial occlusion.

Multilayer dynamic link networks for establishing image point correspondences and visual object recognition

TLDR
A neural network with dynamic links and short-term activity correlations is presented that uses the correspondences for the successful recognition of human faces out of a gallery of 83 independent of their hairstyle and demonstrates the potential for the recognition of objects independently of the background, which was not possible with earlier systems.

Object recognition by dynamic link matching with multiple blob formation

  • H. UmekiH. Mizutani
  • Computer Science
    Proceedings of International Workshop on Neural Networks for Identification, Control, Robotics and Signal/Image Processing
  • 1996
TLDR
A one-to-many object matching system based on neural dynamic link architecture that can roughly detect the matched regions with neighborhood-preserving mappings without global cost functions and extract matched regions from the input layer.

2d object recognition by adaptive feature extraction and dynamical link graph matching

TLDR
A mid-level system for 2D object recognition, comprised of an adaptive feature extractor and a dynamical link graph matcher, that provides the capacity for a richer (than binary) feature set and by increasing the processing speed through improved dynamics and a reduction in the number of nodes needed to represent an object.

Learning invariant object recognition from temporal correlation in a hierarchical network

Internal Report 96{05 Face Recognition by Dynamic Link Matching Face Recognition by Dynamic Link Matching

TLDR
A system for invariant and robust recognition of objects from camera images based on the principles of temporal feature binding and dynamic link matching that aspires both to be a model for biological object vision and to be at the cutting edge of technological achievement.

Simultaneous Recognition of Multiple Objects Using the MEM Model

TLDR
A new self-organizing neural network model, called the Multiple Elastic Modules (MEM), is applied to the task of invariant object recognition through labeled graph matching, where topologically ordered feature labels of an object are matched to the features present in the input scene.

Learning object models from graph templates

TLDR
A novel neural network (NN) based system for detecting rigid objects from their (2-D) gray-level image images and a novel network of NN (NoNN) is proposed to learn the examples of object model graph templates.
...

References

SHOWING 1-10 OF 65 REFERENCES

Size and distortion invariant object recognition by hierarchical graph matching

TLDR
A neural system for invariant object recognition is presented, an application of the dynamic link architecture, which owes its strength to an enhancement of tradition neural networks by a new kind of variable to express the hierarchical grouping of neurons.

Distortion invariant object recognition by matching hierarchically labeled graphs

TLDR
A graph-matching process of object recognition is proposed, applied to face recognition, which amounts to labeled graph matching, with Gabor jets forming labels to nodes and topology determining links.

Pattern recognition by labeled graph matching

Pattern recognition and associative memory as dynamical processes in a synergetic system. I. Translational invariance, selective attention, and decomposition of scenes.

TLDR
It is demonstrated how this model for associative memory and pattern recognition can be cast into a form which is translation invariant and how partially hidden faces in scenes can be recognized by means of the control of attention parameters of specific patterns.

Relations between the statistics of natural images and the response properties of cortical cells.

  • D. Field
  • Computer Science
    Journal of the Optical Society of America. A, Optics and image science
  • 1987
TLDR
The results obtained with six natural images suggest that the orientation and the spatial-frequency tuning of mammalian simple cells are well suited for coding the information in such images if the goal of the code is to convert higher-order redundancy into first- order redundancy.

An evaluation of the two-dimensional Gabor filter model of simple receptive fields in cat striate cortex.

TLDR
It seems that an optimal strategy has evolved for sampling images simultaneously in the 2D spatial and spatial frequency domains and the Gabor function provides a useful and reasonably accurate description of most spatial aspects of simple receptive fields.

A dynamic model for image registration

Constrained Nets for Graph Matching and Other Quadratic Assignment Problems

TLDR
The main point of the elastic net algorithm is seen to be in the way one deals with the constraints when evaluating the effective cost function (free energy in the thermodynamic analogy), and not in its geometric foundation emphasized originally by Durbin and Willshaw.

Modeling perceptual grouping and figure-ground segregation by means of active reentrant connections.

TLDR
This model is modified and extended to address the problems of perceptual grouping and figure-ground segregation in vision and is able to link the elements corresponding to a coherent figure and to segregate them from the background or from another figure in a way that is consistent with the Gestalt laws.

The Correlation Theory of Brain Function

TLDR
A summary of brain theory is given so far as it is contained within the framework of Localization Theory, according to which synaptic modulation introduces flexibility into cerebral networks which is necessary to solve the invariance problem.
...