Quantum-inspired immune clone algorithm and multiscale Bandelet based image representation

  • Authors:
  • Shuyuan Yang;Min Wang;Licheng Jiao

  • Affiliations:
  • Department of Electrical Engineering, Institute of Intelligent Information Processing, Xidian University, Xi'an, 710071, China;Department of Electrical Engineering, National Key Laboratory of Radar Signal Processing, Xidian University, Xi'an, 710071, China;Department of Electrical Engineering, National Key Laboratory of Radar Signal Processing, Xidian University, Xi'an, 710071, China

  • Venue:
  • Pattern Recognition Letters
  • Year:
  • 2010

Quantified Score

Hi-index 0.10

Visualization

Abstract

In this letter, we proposed a quantum-inspired immune clone algorithm (QICA) and wavelet packet (WP) based adaptive multiscale Bandelet (QICAWP-AMB) for image representation. Multilevel wavelet packet decomposition is performed on the images and then dyadic partition is employed to obtain some squares on which to perform Bandeletization. Moreover, in order to locate the optimal direction in each square, a new cost function independent with the threshold T is proposed, and QICA is employed to obtain more accurate geometric flows. Experiments on image coding are taken to compare the efficiency of our proposed method with that of Wavelets, the second generation Bandelet, JPEG2000, SPIHT, GAWP (genetic algorithm and wavelet packet) and ICAWP (immune clone algorithm and wavelet packet) based methods, and the superiorities of QICAWP-AMB to its counterparts can be observed in both peak signal to noise (PSNR) and edge preservation index (EPI) at low bit rate compression.