Related Experiment Video
Updated: Jul 7, 2026
11:41
Automated Sholl Analysis of Digitized Neuronal Morphology at Multiple Scales
Published on: November 14, 2010
Generating and coding of fractal graphs by neural network and mathematical morphology methods
IEEE Transactions on Neural Networks
|January 1, 1996
Summary
We developed a fractal graph generation algorithm using probabilistic logic neuron networks. This method provides optimal, compressed encoding for these self-similar graphs, potentially aiding image compression and pattern recognition.
Area of Science:
- Graph theory
- Artificial intelligence
- Computer science
Background:
- Self-similar (fractal) graphs are complex structures with applications in various fields.
- Efficient representation and recognition of these graphs are challenging.
Purpose of the Study:
- To present an algorithm for generating fractal graphs using probabilistic logic neuron networks.
- To develop a method for compressed encoding and recognition of these graphs.
- To explore alternative generation methods and potential applications.
Main Methods:
- Utilized simple probabilistic logic neuron networks for fractal graph generation.
- Developed an algorithm for finding compressed encoding of the generated graphs.
- Investigated mathematical morphology as an alternative generation technique.
Main Results:
- Successfully generated a class of self-similar (fractal) graphs.
- Demonstrated that these graphs can be represented by a set of compressed encoding.
- Showed that the encoding is optimal (minimal length).
- Confirmed that the same graphs can be generated via mathematical morphology.
Conclusions:
- The proposed algorithm efficiently generates and encodes fractal graphs.
- The findings suggest potential applications in image compression and pattern recognition.
- This work bridges graph theory, neural networks, and mathematical morphology.
