Abstract

The objective of this thesis is to acquire abstract image features through statistical modelling in the wavelet domain and then based on the extracted image features, develop an effective content-based image retreival (CBIR) system and a fragile watermarking scheme. In this thesis, we first present a statistical modelling of images in the wavelet domain through a Gaussian mixture model (GMM) and a generalized Gaussian mixture model (GGMM). An Expectation Maximization (EM) algorithm is developed to help estimate the model parameters. A novel similarity measure based on the Kullback-Leibler divergence is also developed to calculate the distance of two distinct model distributions. We then apply the statistical modelling to two application areas: image retrieval and fragile watermarking. In image retrieval, the model parameters are employed as image features to compose the indexing feature space, while the feature distance of two compared images is computed using the novel similarity measure. The new image retrieval method has a better retrieval performance than most conventional methods. In fragile watermarking, the model parameters are utilized for the watermark embedding. The new watermarking scheme achieves a virtually imperceptible embedding of watermarks because it modifies only a few image data and embeds watermarks at image texture edges. A multiscale embedding of fragile watermarks is given to enhance the embeddability rate and on the other hand, to constitute a semi-fragile approach.

Highlights

  • 1.1 M otivation and O b jectivesMany image applications require an accurate modelling of images to have a better under­ standing and utilization of image contents

  • In C hapter 3, we develop a statistical modelling of images in the wavelet domain through Gaussian mixture model (GMM) and generalized Gaussian mixture model (GGMM)

  • Many image applications are based on the image characteristics obtained from image mod­ elling

Read more

Summary

M otivation and O b jectives

Many image applications require an accurate modelling of images to have a better under­ standing and utilization of image contents. Our interest lies in the derivation of some new feature extraction methods based on the developed statistical modelling to have a compact indexing feature space. It is desired th a t the 2 extracted image features can reflect some characteristics of human perception and have a sound image retrieval performance. Most traditional fragile watermarking methods inevitably need to modify a large amount of image data for watermark embedding in order to protect the whole image area. We aim at developing a new fragile watermarking method th at modifies a few image data for watermark embedding, while keeping sensitive to any image tampering.

Background
Im age R etrieval
Im age W aterm arking
Watermark Decoding
C ontribution
Structure of Thesis
The EM A lgorithm for M ixture M odels
M clximum L ikelihood
T he G eneral EM A lgorithm
T h e K ullback-Leibler D ivergence
A G aussian M ixture M odel in th e W avelet D om ain
A n EM A lgorithm for th e G aussian M ixture M odel
G eneral M inkow ski D istances
30 Assume we have two Gaussian mixture distributions
O verview of the P roposed C B IR System
The Indexing Feature Space for Image R etrieval
K ullback D ivergence Based Sim ilarity M easure
Sim ulation R esults
37 D bl Filters Db2 Filters
Sum mary
E m bedding Inform ation into the S tatistical M odel
M ultiscale Em bedding of A uthentication M essages
Sum m ary
Findings
Conclusions and Future Work
Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call