Related Experiment Videos
A New Neural Model for Invariant Pattern Recognition
Shuenn Shyang Wang1, Wen Gou Lin
1Department of Electrical Engineering, Tatung Institute of Technology, 40 Chungshan North Road, 3rd Sec., Teipei, Taiwan, People's Republic of China
Summary
This study introduces a novel neural model for invariant pattern recognition, enabling accurate identification despite variations in position, rotation, and scale. The proposed model effectively handles these transformations for robust pattern recognition applications.
Area of Science:
- Computer Science
- Artificial Intelligence
- Machine Learning
Background:
- Pattern recognition systems often struggle with variations in object position, rotation, and scale.
- Achieving invariant pattern recognition is crucial for many real-world applications.
Purpose of the Study:
- To propose a novel neural model for invariant pattern recognition.
- To develop a model capable of recognizing patterns irrespective of their position, rotation, and scale.
Main Methods:
- A cascade connection of four two-dimensional neural network layers was designed.
- The first three layers handle position normalization, rotation normalization, and feature extraction.
- The final layer performs recognition and scale normalization using scale-invariant output neurons.
Main Results:
- Simulation results demonstrate the model's effectiveness.
- The proposed neural model achieves invariant pattern recognition.
Conclusions:
- The developed neural model is simple and effective for invariant pattern recognition.
- The model successfully addresses challenges posed by variations in pattern orientation and size.