This article reports on the development of a new digital watermarking technique for images that are sensitive to blocking artifacts.
Experimental results show that the proposed MDCT based approach produces highly imperceptible watermarked images and is robust to attacks such as compression, noise, filtering, and geometric transformations. The proposed MDCT watermarking technique is applied to fingerprints for ensuring security. The face image and demographic text data of an individual are used as multiple watermarks. An AFIS system was used to quantitatively evaluate the matching performance of the MDCT-based watermarked fingerprint. The high fingerprint matching scores show that the MDCT approach is resilient to blocking artifacts. The quality of the extracted face and extracted text images was computed using two human visual system metrics and the results show that the image quality was high.
810 Seventh Street NW, Washington, DC 20531, United States
Appears in International Journal of Signal Processing (2007), Volume 1 Number 3