Images have to be compressed in order to store and transmit them efficiently. In this paper, we propose a hybrid image compression technique which uses bandlet transform and type-2 fuzzy thresholding. Bandlet transform is used to efficiently represent the anisotropic regularity of edge structures in images and fuzzy thresholding retains the important coefficients based on the measure of ultrafuzziness. The thresholded bandlet transform coefficients are then compressed using a variable length coding technique, viz., arithmetic coding. The proposed method is applied to two-dimensional (2D) images. The results are compared with type-2 fuzzy thresholded wavelet compression. Test results show that the proposed method using bandlet transform gives better results in terms of Peak-Signal-to-Noise Ratio (PSNR) and compression ratio compared to wavelet transform and type-2 fuzzy thresholding.

bandlet transform, type-2 fuzzy thresholding, compression.