Novel Wavelet Domain Based Adaptive Thresholding Using Bat Algorithm for Image Compression

Image compression is the significant study in the arena of image processing owing to its enormous usages and its ability to reduce the storage prerequisite and communication bandwidth. Thresholding is a kind of image compression in which computational time increases for multilevel thresholding and hence optimization techniques are applied. The quality of reconstructed image is superior when discrete wavelet transform based thresholding is used as compared to when it is not applied. Both particle swarm optimization and fire fly algorithm becomes unstable when the velocity of the particle becomes maximum and when there is no bright firefly in the search space respectively. To overcome the above-mentioned drawbacks bat algorithm based thresholding in frequency domain is proposed. Echolocation is the sort of sonar used by micro-bats. The way they throng their prey, overcoming the hurdles they come across, pinpointing nestling gaps have become the main motivation research in artificial intelligence. With the feature of frequency tuning and having the benefit of automatic zooming, bat algorithm produces superior PSNR values and quality in reconstructed image and also results in fast convergence rate as compared to state of art of optimization techniques.