Gaussian Base Image Data Embedding for Scalable Fragile Compression Attack

Project Discription

The objective of watermarking multimedia is to prevent unauthorized copies from being distributed without the consent of the original owner. The invisible watermarking algorithm proposed should be robust to common watermark attacks. Such a watermarking algorithm can be applied to digital audio, video, or image files. Extracting the watermark from any of these three media unambiguously identifies the head (owner), thereby creating the counterfeit of the main digitized medium extra challenging.  In this work a new approach of watermarking is studied in detail. Based on human view, Low frequency band of the image is use, so here it make a invisible watermarking technique base on Gaussian function robustness of image is improve. Results shows that the proposed work is producing the results which maintain the image quality as well as robustness against the fragile images. Watermark obtain from the extraction method is having 0.24% of BER which is quite impressive results for all the researchers. In future work can be improve for other attacks such as geometry of image.

Embedding

 In this step selected coefficient least significant values are replace by Gaussian random values. 

Extraction

 In case of extraction all steps remain same no need of any other calculation as above xor operation will recalculate the pixel value to its original form as done in above steps. But this is possible only in case of no attack condition while in case of attack those pixel which are affected get near by values.

Project Sample Image