Image Compression Based on Spectral Graph Wavelet Transform

Authors

OMANI Katia
Laboratoire d’Analyse et Modélisation des Phénomènes Aléatoires (LAMPA)Mouloud Mammeri University of Tizi-Ouzou(UMMTO).Tizi-ouzou
CHERIFI Mehdi
Laboratoire d’Analyse et Modélisation des Phénomènes Aléatoires (LAMPA)Mouloud Mammeri University of Tizi-Ouzou(UMMTO).Tizi-ouzou
LAHDIR Mourad
Laboratoire d’Analyse et Modélisation des Phénomènes Aléatoires (LAMPA)Mouloud Mammeri University of Tizi-Ouzou(UMMTO).Tizi-ouzou

Synopsis

A new approach for image compression based on spectral graph wavelet transform (SGWT) is presented in this work. By converting the image into the frequency domain with the SGWT, the resulting frequency components can be quantized to eliminate the visually insignificant data, thereby decreasing the amount of information required for storage. Entropy and Huffman coding are then employed to exploit the redundant properties of the quantized frequency samples, resulting in a compressed representation of the image. The entire process is reversible, enabling the reconstruction of the original image from the compressed form. Our approach was applied to gray scale and colored images, and the results of the experiments were encouraging.

ICAECE2023
Published
February 5, 2024