Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/6986
標題: 利用小波轉換及向量化編碼法之影像隱藏技術
Image Hiding Scheme based on Vector Quantization and Wavelet Transformation
作者: 黃育銓
Huang, Yu-Chuan
關鍵字: Data Hiding;資料隱藏;Cover Image;Secret Image;Stego Image;Vector Quantization;Discrete Wavelet Transform;Least Significant Bit;掩護影像;機密影像;偽裝影像;向量化編碼法;離散小波轉換;最低位元法
出版社: 電機工程學系
摘要: 
本論文主要探討三張大小相同之灰階影像,機密影像(Secret Image)、掩護影像(Cover-Image)、及偽裝影像(Stego-Image)。本文主要目的是希望能夠將一張機密影像隱藏至一張掩護影像後產生一張所謂偽裝影像,偽裝影像與掩護影像幾近相同,因此機密影像於網路上傳遞將不易被察覺甚至破壞。即使,一旦掩護影像被有心人士竄改、壓縮、或裁切破壞等,依據本文之技術仍可以取出機密影像。
在我們所提出之資料隱藏(data-hiding)技術裏,首先為了提高隱藏的資料量,機密影像會先經過PNN、LBG、及字碼簿(codebook)排序等向量編碼(Vector Quantization)方法做資料壓縮。接者,為了增強抗壓縮之強韌性,掩護影像會先經過三層之離散小波轉換(Discrete Wavelet Transform)後,再將索引值隱藏在小波係數裏。最後,再經由反離散小波轉換(Inverse Discrete Wavelet Transform)後產生一張具有強韌性的偽裝影像。依本文的技巧,我們所重建的機密影像當JPEG壓縮品質在百分之十一以下的壓縮率,其影像品質完全不受影響,甚至在百分之五十的壓縮率,其影像仍可清晰辨識。

This dissertation primarily explores three grey-level images of the same size 256256 pixels, the secret image, the cover-image, and the stego-image. The major goal of this dissertation is to embed a secret image into a cover-image to constitute a so-called stego-image such that the stego-image is visually indistinguishable from the original cover-image. Hence, the interceptors will not notice the existence of secret data in the stego-image. Furthermore, through our proposed data hiding scheme, the secret image still can be extracted from the stego-image even when the stego-image was modified or destroyed by popular image processing techniques, such as cropping and image compression.
In our proposed scheme. First, in order to enhance the hiding capacity, the secret image was compressed by those techniques of vector quantization, PNN、LBG、and codebook sorting. Next, for the purpose of robust property against compression, the indices were embedded into the wavelet coefficients of the cover-image which was processed through the three-level discrete wavelet transform. Finally, a robust stego-image was obtained through the processing of inverse discrete wavelet transform. According to our experimental results, the reconstructed secret image can keep its original quality against JPEG compression with compression ratio below 11% and is still recognizable while the compression ratio is up to 50%.
URI: http://hdl.handle.net/11455/6986
Appears in Collections:電機工程學系所

Show full item record
 

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.