الملخص

Abstract Geometrical shapes are so important in pattern and image recognition So many research use proposed in this direction . In this paper study basic geometrical shapes [circle, triangle, rectangle,…,etc] are studied and small database was build for them based on their eigenvalue. Eigenvalues and eigenvectors studied in this paper as shape descriptor parameters, which found it can be strongly used due to the high difference shapes. backprogation neural network was achieved to a speed path recognition process.