A novel hybrid image compression technique: Wavelet-MFOCPN

Abstract: In this paper a novel hybrid image compression technique is proposed. This technique inherits the properties of localizing the global spatial and frequency correlation from wavelets and classification and functional approximation tasks from modified forward-only counterpropagation neural network (MFO-CPN) for image compression. Several benchmark test images are used to investigate usefulness of the proposed technique. Experimental results of the proposed technique showed an enhancement in performance measures with respect to decoded picture quality and compression ratios, compared to the existing wavelet and neural network based image compression technique