|
|
|
Handwritten Character Recognition (ICR)
| |
The Concept
|
|
|
ICR recognizes characters using the following steps:
- Training phase — off line
- Find topology.
- Calculate features.
- For each topology train a neural network.
- Recognition phase
- Find topology.
- Calculate features.
- Use a neural network that was trained on this topology,
and produce a probability for each class.
Our ICR uses two methods to define topologies:
- Horizontal/Vertical lines
Each character is segmented to horizontal and vertical lines.
- Primitives shapes
Each character is segmented into loops, lines, curves, etc.
For each method there is a different set of neural networks.
Voting is carried out between the results of these two methods.
The Algorithm Scheme
|
|
|
Examples
|
|
|
Samples of recognized images from NIST database.
|
|
|
|
|
