الفهرس 
chapter 1Only 14 pages are availabe for public view
 |  | from | 149 |  |  |
| | | from | 149 | | |
Abstract
Three new algorithms are proposed in the area of understanding and analyzing shapes. The first algorithm is an effective, low computational cost technique to find the orientation of shapes that have several axes of symmetry. A simple method to calculate the average angle of the shape’s axes of symmetry is defined using only the first moment of inertia to reduce the computational cost. The second algorithm is a novel rotation-invariant neural-based pattern recognition system. A new image preprocessing technique is defined to extract rotation-invariant descriptive patterns from the shapes to be used in both the neural network training and application phases to cancel any rotation effect. The third algorithm is a new technique to measure the orientability of different shapes. The proposed technique tends to use the simplicity of the ellipse orientability to obtain orientability of the other shapes. Results reveal that the proposed techniques are very effective with low computational cost.