The present study proposes a novel technique for copyright protection by utilizing digital watermarking of Images. The watermark is embedded and detected by using Functional Link Artificial Neural Network (FLANN) and Discrete Cosine Transform (DCT). The exhaustive simulation results of the proposed scheme show improved performance over the existing methods in all cases, i.e. when the watermarked image is subjected to compression, cropping, sharpening, blurring and noise. Comparative analysis with an existing neural approach shows the superiority of the proposed scheme of computational complexity and performance.

No abstract

A new scheme for Arabic handwriting recognition by means of mat lab software is proposed. It can be seen as construction of neural networks that associate the hand writing Arabic text with the stored Arabic text in a specific data base are nearly equivalent. This result indicates that the new scheme have the potential to be robust in the presence of noisy data input

No abstract