The application of Terahertz time domain spectral analysis and image hashing based on wave atom transform

Abstract

Wave atom transform is a new technology which has been introduced into signal processing and image processing in recent years. Due to its special sparser expansion for the oscillatory functions, it has a very promising application prospect. In this thesis, the wave atom transform is embedded into the Terahertz time domain spectral analysis and image hashing. In the Terahertz time domain spectral analysis, reflection happens when the Terahertz wave goes through the surface of the substance. Therefore, the traditional Fourier Transform analysis or other transforms analysis like wavelet and curvelet have limitations in providing clear spectrum characteristics of the wave as the basis models do not imitate the reflection behavior of the wave. Compared with conventional methods, the wave atom transform is expected to have a better performance when dealing with wave reflections and transmission phenomena. The experiment results demonstrate that the proposed method can enlarge the resolution in the frequency domain and a number of covered significant spectrums can be refined. The results of our work show that the proposed method is the most suitable for the spectrums which have one or two distinct wave crests. For spectrums which are monotonic decreasing, i.e. absorption characteristic is inversely related to frequency or no wave reflection occurs, the improvements are not significant. Overall, when compared with the conventional analysis methods, the proposed method achieves a better spectrum performance by showing up very closely located amalgamated peaks in the Fourier spectrum for waves with reflection behavior. In the image hashing, a new strong image hashing scheme is proposed based on wave atom transform. The scheme can better authenticate images by precisely distinguishing the maliciously tampered images from the non-maliciously processed ones. For better detection sensitivity, gray code is applied instead of natural binary code to optimize the Hamming distance. Randomizations are also performed using a Rényi chaotic map for the purposes of secure image hashing and key sensitivity. The experiment results show that the proposed scheme is robust to non-malicious content-preserving operations and also fragile to malicious content-altering operations. The scheme also outperforms DCT and DWT based schemes in terms of performance characteristic curves.

Date of Award	15 Feb 2013
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Lee Ming CHENG (Supervisor) & Hong YAN (Co-supervisor)