Fingerprint Image Compression using Sparse Representation and Enhancement with Wiener2 Filter

Abstract

Abstract:A technique for enhancing decompressed fingerprint image using Wiener2 filter is proposed. First compression is done by sparse representation. Compression of fingerprint is necessary for reducing the memory consumption and efficient transfer of fingerprint images. This is very essential for the application which includes access control and forensics. So the fingerprint image is compressed using sparse representation. In this technique, first dictionary is constructed for patches of fingerprint images. Then a fingerprint is selected and the coefficients are obtained and encoded. Thus the compressed fingerprint is obtained. But when the fingerprint is reconstructed, it is affected by noise. So Wiener2 filter is used to filter the noise in the image. The ridge and bifurcation count is extracted from decompressed and enhanced fingerprints. The experiment result shows that the enhanced fingerprint image preserves more bifurcation than decompressed fingerprint image. The future analysis can be considered for preserving ridges.

Discrete Shearlet Transform and Lempel-Ziv Welch Coding for Lossless Fingerprint Image Compression

“…Lossless fingerprint image compression using Huffman coding, arithmetic coding for lossless compression and discrete transforms such as discrete wavelet transform, discrete cosine transform and discrete shearlet transform is made by Kadim et al (2020), A new technique for multispectral fingerprint biometric systems based on image compression and employing wavelet decomposition and Huffman coding has been introduced by Sharma et al (2020), fingerprint image compression using sparse representation and enhancement with Wiener2 Filter is a method developed by Joseph and Joseph (2015). Radhika et al (2022) made the fingerprint compression algorithm based on singular value decomposition in sparse representation; (Murthy et al, 2022) made the fingerprint compression algorithm based on sparse representation and (Saikrishna and Sreenivasulu, 2015) made the fingerprint compression algorithm based on singular value decomposition in sparse representation.…”