Gray-Scale Fingerprint Image Compression Based on the Hybrid-NRCT

As a directional multiresolution image representation, the contourlet transform can efficiently capture curved and oriented geometrical structures in images. However, the contourlet transform has the drawback of a 4/3 redundancy in its oversampling ratio. Recently, Zhang and Moloney have developed a nonredundant version of the contourlet transform, called the nonredundant contourlet transform (NRCT) and have demonstrated that this 4/3 redundancy can be eliminated by the NRCT. With the advantages of critical sampling and perfect reconstruction, the NRCT is suitable for tracking and efficiently coding oriented structure in images, such as the texture of ridges in fingerprint images. Moreover, as an extension of the wavelet transform, the NRCT is easily compatible with the wavelet transform. A new transform which combines the NRCT with the wavelet transform is called the hybrid-NRCT. This paper proposes a compression scheme for fingerprint images using the hybrid-NRCT, and compares its performance with other transform-based fingerprint image compression schemes.